Article

Use of a digital camera as alternative method for non-destructive detection of the leaf chlorophyll content and the nitrogen nutrition status in wheat

Authors:
To read the full-text of this research, you can request a copy directly from the authors.

Abstract

In this paper, the use of a digital consumer camera for the non-destructive detection of the N nutritional status is compared with two alternative methods, namely SPAD and reflectance spectrometry in three field experiments. The image analysis method consisted of segmentation and successive analysis of the foreground color, i.e. only green plant parts. Thus, also analysis of canopies with small degree of ground cover is possible. All methods gave comparable results, while the effort necessary was considerably higher when using the chlorophyll meter. With spectral measurements, the biomass and leaf nitrogen content could not be clearly differentiated; chlorophyll measurements do not reflect biomass, whereas the described procedure of image analysis permits the consideration both. If used properly, digital image analysis is a valuable tool for the determination of the N nutrition status under field conditions, with low costs and labor requirements.

No full-text available

Request Full-text Paper PDF

To read the full-text of this research,
you can request a copy directly from the authors.

... CropCircle sensor (Holland Scientific, Lincoln, 72 NE) was applied to measure the canopy reflectance of cucumber (Padilla et al., 2017) and muskmelon (Padilla et al., 2014) to predict N status. Spectroscopy and digital or multispectral cameras were also available and easy methods (using different filters) to evaluate the N content of plants widely (Baresel et al., 2017;Yang et al., 2008). Since chlorophyll is a criterion to show the appropriate supply of nitrogen, a chlorophyll meter (Soil-Plant Analyses Development, SPAD) as a reliable sensor was applied to estimate leaf greenness and N content of different crops, including rice (Larijani and Farokhi-Teymorlou, 2012), muskmelon (Padilla et al., 2014), and maize (Schmidt et al., 2011). ...
... Then, the treatments were separated from each other by receiving different percentages of 46% urea fertilizer with weekly repetition (Table 2). Gianquinto et al. (2011) and Padilla et al. (2014and 2017 suggested five fertilizer treatments in their studies on cucumbers. ...
Article
Full-text available
Monitoring physiological parameters of plant and fertilizer requirements are the basic principles in precision farming. Non-destructive and accurate remote sensors make this process feasible and light-handed. The present study evaluates the efficiency of GreenSeeker (GS) and Soil-Plant Analyses Development (SPAD) in Nitrogen (N) fertilizer management. Normalized difference vegetation index (NDVI) was measured with GS and compared to SPAD. Fertigation with 5 different N treatments was applied to 100 pots. The first treatment (N1) had no N concentration, while the other treatments (i.e., N2, N3, N4, and N5) received 2, 9, 22, and 37 mmol.L-1 weekly, respectively. Next, the effect of volumetric fertilizing was investigated by adding supplemental fertilizer to N1 to N3 pots 71 days after planting. Nitrogen concentration in the leaf and first growing fruit was tested using the Kjeldahl method. The results of applied sensors confirmed with visible-near infrared spectroscopy at 200-1100 nm wavelength. NDVI, soil-adjusted vegetation index, and chlorophyll index were calculated from the available spectra and compared to the sensor outputs. Strong correlations were obtained between NDVI and all indices derived from spectra, especially in the vegetative phase. The results showed a strong correlation of NDVI with N rate, especially after supplemental fertilizing. Since the vegetation indices from spectra almost correlated well with NDVI and SPAD in all treatments, spectroscopy monitoring of cucumber could be a precise alternative technique. Linear and nonlinear regressions were applied to model variations of NDVI and SPAD. This study demonstrated the feasibility of using GS for N management according to its sensitivity to cucumber N status.
... The color features extracted from the digital images have been successfully used to monitor the leaf chlorophyll content or relative index of crops or trees [29,47,48]. These phenotypic features can also reflect the morphological characteristics of plants, i.e., leaf area index and shoot dry weight of rice [49] and aboveground biomass of wheat [25]. ...
... Although there were significant relationships of H with plant weight and leaf area and of b and 2G-R-B with the leaf number on main stem when the images were taken from the top view, the color parameters did not reflect the morphological characteristics in general, especially from the side view. Previous studies pointed out that the shooting angles of images caused differences in reflected and refracted light, affecting the value of the acquired parameters, and the color parameters extracted from images from the top vertical view had better correlations with leaf color and the morphological index compared with images from other views [25,48]. ...
Article
Full-text available
The accurate and efficient screening of waterlogging-tolerant cultivars is an effective way to mitigate waterlogging damages. An experiment was conducted to evaluate the performance of 28 wheat varieties mainly planted in the middle and lower reaches of the Yangtze River, China, under control and waterlogging conditions. When the 15-day waterlogging that was initiated at the third-leaf stage was completed, the aboveground dry weight, plant height, leaf number on main stem, culm number, leaf area, and SPAD readings of wheat seedlings were significantly decreased by 14%, 11%, 6%, 13%, 14%, and 15% compared with the control treatment (maintaining approximately 80% of field capacity), respectively. The results showed that the percentage reductions in the dry weight and leaf area under stress accurately represented the influence of the majority of the measured agronomic traits and were significantly negatively correlated with the respective dry weight and leaf area of different cultivars under waterlogging. This suggests that dry weight and leaf area can be used as agronomic traits for screening waterlogging-tolerant cultivars. The comprehensive evaluation value of waterlogging tolerance (CEVW) was closely related to the percentage reduction in dry weight, plant height, culm number, leaf area, and SPAD reading. The range of CEVW was 0.187–0.819, indicating a wide variation in the waterlogging tolerance of the wheat cultivars. Comparing the top-view images, the phenotypic texture parameters (dissimilarity, homogeneity, and angular second moment (ASM)) extracted from the side-view images better reflected the dry weight, plant height, and leaf area under different water treatments. The percentage reduction in ASM had the strongest correlation with CEVW (root mean square error = 0.109); thus, the ASM is recommended as a suitable phenotypic parameter to evaluate waterlogging tolerance. The present results provide references for the rapid and intelligent screening of waterlogging-tolerant wheat cultivars, but future studies need to consider the stress evaluation of the adult plants.
... Drought resistance (DR) is defined as the mechanism causing minimum water loss in a water deficit environment while maintaining its production. DR is determined by how quickly and efficiently a plant senses changing environmental conditions, and how the plant adopts and combines the aforementioned strategies in response to diminished water availability (Baresel et al., 2017). DR is linked to a combination of morphological, anatomical and physiological traits (Lozano et al., 2020). ...
... Nitrogen supply has a strong influence on leaf growth (He and Dijkstra, 2014). Plant leaf area growth promotes photosynthesis at the same time (Baresel et al., 2017). Chlorophyll is the main product of photosynthesis. ...
Article
Full-text available
Drought is a climatic event that considerably impacts plant growth, reproduction and productivity. Toona sinensis is a tree species with high economic, edible and medicinal value, and has drought resistance. Thus, the objective of this study was to dynamically monitor the physiological indicators of T. sinensis in real time to ensure the selection of drought-resistant varieties of T. sinensis. In this study, we used near-infrared spectroscopy as a high-throughput method along with five preprocessing methods combined with four variable selection approaches to establish a cross-validated partial least squares regression model to establish the relationship between the near infrared reflectance spectroscopy (NIRS) spectrum and physiological characteristics (i.e., chlorophyll content and nitrogen content) of T. sinensis leaves. We also tested optimal model prediction for the dynamic changes in T. sinensis chlorophyll and nitrogen content under five separate watering regimes to mimic non-destructive and dynamic detection of plant leaf physiological changes. Among them, the accuracy of the chlorophyll content prediction model was as high as 72%, with root mean square error (RMSE) of 0.25, and the RPD index above 2.26. Ideal nitrogen content prediction model should have R² of 0.63, with RMSE of 0.87, and the RPD index of 1.12. The results showed that the PLSR model has a good prediction effect. Overall, under diverse drought stress treatments, the chlorophyll content of T. sinensis leaves showed a decreasing trend over time. Furthermore, the chlorophyll content was the most stable under the 75% field capacity treatment. However, the nitrogen content of the plant leaves was found to have a different and variable trend, with the greatest drop in content under the 10% field capacity treatment. This study showed that NIRS has great potential for analyzing chlorophyll nitrogen and other elements in plant leaf tissues in non-destructive dynamic monitoring.
... Remote sensing (e.g., Iglhaut et al. 2019;Puletti et al. 2019) or leaf spectroscopy (e.g., Colombo et al. 2012;Cotrozzi and Couture 2019;Jacquemoud and Ustin 2019) have great potential to address these tasks (Humplík et al. 2015;Mahlein et al. 2018). Moreover, reliable, fast, and user-friendly monitoring systems at affordable costs would facilitate the spread of breakthrough technology, particularly by remote and "near-surface" sensing techniques (Baresel et al. 2017;Di Gennaro et al. 2018). Recently, near-surface sensing and high pixel resolution imagery allowed the targeting of single plant parts (e.g., leaf, stem or canopy by segmentation techniques, Perez-Sanz et al. 2017;Chianucci et al. 2019). ...
... In Table 1, selected 'dcraw' parameters are reported to capture scene full brightness range at maximum radiometric resolution (16 bit) to preserve the linear dependence of digital numbers (DN) to actual brightness (Chianucci et al. 2019). A segmentation procedure was applied to separate the leaf blade image from the background (Baresel et al. 2017) and leaf reflectance was calculated as DN leaf, x / DN reference, x , where x was G, R, and NIR band, respectively, and DN was the mean digital number either over the whole leaf blade or the white reference. The following vegetation indices (VIs) were calculated-namely, normalized difference vegetation index, NDVI = (NIR − Red)/(NIR + Red), (Rouse et al. 1973); green normalized difference vegetation index, GNDVI = (NIR − Green)/(NIR + Green), (Gitelson et al. 1996), which is also equivalent to minus the normalized difference water index (NDWI, McFeeters 1996) and near-infrared reflectance of vegetation, NIRv = NDVI · NIR (Badgley et al. 2017), an index proposed as a proxy of light-saturated photosynthesis (A max ) at leaf scale for deciduous tree species included Quercus (Wong et al. 2020). ...
Article
Full-text available
A user-friendly and affordable broad-band digital Near Infrared (NIR) camera (Canon Pow-erShot S110 NIR) was compared with a narrow-band reflectance spectrometer (USB2000, Ocean Optics) at leaf scale for monitoring changes in response to drought of three ecologically contrasting Quercus species (Q. robur, Q. pubescens, and Q. ilex). We aimed to (a) compare vegetation indices (VIs; that is: NDVI, Normalized Difference Vegetation Index; GNDVI, Green NDVI and NIRv, near-infrared reflectance of vegetation) retrieved by NIR-camera and spectrometer in order to test the reliability of a simple, low-cost, and rapid setup for widespread field applications; (b) to assess if NIR-camera VIs might be used to quantify water stress in oak seedlings; and (c) to track changes in leaf chlorophyll content. The study was carried out during a water stress test on 1-year-old seedlings in a greenhouse. The camera detected plant status in response to drought with results highly comparable to the visible/NIR (VIS/NIR) spectrometer (by calibration and standard geometry). Consistency between VIs and morpho-physiological traits was higher in Q. robur, the most drought-sensitive among the three species. Chlorophyll content was estimated with a high goodness-of-fit by VIs or reflectance bands in the visible range. Overall, NDVI performed better than GNDVI and NIRv, and VIs performed better than single bands. Looking forward, NIR-camera VIs are adequate for the early monitoring of drought stress in oak seedlings (or small trees) in the post-planting phase or in nursery settings, thus offering a new, reliable alternative for when costs are crucial, such as in the context of restoration programs.
... Data plots present a logarithmic decline with these two colors when CHL content increases [8], which indicates a lower sensitivity of the R and G models under higher pigment content. Other color indices such as R-B, G-B, R+B, R+G, B+G, R+G+B, R/B and G/B, show a good correlation with chlorophyll content [9]. These indices measure spatial and temporal plant photosynthetic activity variations. ...
... Measurement of chlorophyll on land is also done using Unmanned Aerial Vehicles (UAVs) [7]. Imagery using a hyperspectral camera can be used to do chlorophyll estimation [8]. A spectrometer was used to estimate the chlorophyll content in rice plants [9]. ...
Conference Paper
Leaves can be used to assess the health of a plant. The chlorophyll content and nitrogen status of plants might be used to assess plant health. Chlorophyll content of plant leaves is measurable in the laboratory using dimethyl sulfoxide or the Kjeldahl technique. These destructive investigations of plant tissues accurately evaluates chlorophyll concentration and nitrogen status. However, this procedure is costly and time-consuming. Using a SPAD instrument, non-destructive chlorophyll measurements were performed. However, the price of the instrument is high. This research intends to develop a technique for detecting chlorophyll content using a smartphone's built-in camera. The approach used is an advancement of ground-based remote sensing technology. The object of this research is maize. An RGB camera on a mobile phone was used to gather data on the greenness level of the leaf. Concurrently, the SPAD value was measured to determine the chlorophyll content. In this study, there were three primary phases, capturing leaf images with geotags, leaf image processing, and vegetation index (VIs) analysis. This research demonstrated that a mobile camera has potential as a tool for measuring chlorophyll content depending on the level of leaf greenness. Additionally, the GPS function on each data provides the chlorophyll distribution in a field.
... Assessing the N status in plants is essential for maximizing the efficiency of nitrogen fertilizers. Traditional methods involving destructive chemical analyses of plant samples are highly accurate but impractical for large-scale assessments due to the need for extensive sampling and expensive laboratory procedures (Baresel et al., 2017;Frels et al., 2018). To overcome these limitations, non-destructive spectral analyses have emerged as a valuable tool for crop monitoring and assessing plant N status without the need for intensive sampling (Verrelst et al., 2015;Elsayed et al., 2018;Prey and Schmidhalter, 2019;Sahoo et al., 2023a). ...
Article
Full-text available
Nitrogen responses vary under diverse agronomic management practices, influencing vegetation indices (VIs) and productivity across different ecological conditions. However, the proper quantification of these responses under various crop establishment methods with varied nitrogen levels is rarely studied. Therefore, a field experiment was conducted to investigate the impact of varying nitrogen levels on VIs, growth parameters, yield attributes, yield, and economic aspects of transplanted rice (TR) and direct-seeded rice (DSR). The experiment was conducted in the randomized block design consisted seven N levels, which included 0% recommended dose of nitrogen (RDN) or no nitrogen (N0), 33.33% RDN (N1), 66.66% RDN (N2), 100% RDN (N3), 133.33% RDN (N4), 166.66% RDN (N5) and 200% RDN (N6), and replicated thrice. The plots with higher N levels demonstrated increased values of VIs and treatment N3 (120 kg N ha⁻¹), N4 (160 kg N ha⁻¹), N5 (200 kg N ha⁻¹), and N6 (240 kg N ha⁻¹) showed no statistically significant differences in NDVI (normalized difference vegetation index), RVI (ratio vegetation index), NDRE (normalized difference red edge), and GNDVI (green normalized difference vegetation index) values across the various growth stages of rice. The application of treatment N4 resulted in the highest number of panicles m⁻² (348.2 in TR, 376.8 in DSR), filled grains panicle⁻¹ (74.55 in TR, 62.43 in DSR), and a 1,000-grain weight of 26.92 g in TR and 26.76 g in DSR. The maximum yield (4.89 t ha⁻¹) was obtained in transplanted rice at treatment N4 and, 8.15% yield reduction was noted in DSR for the same treatment, which was statistically equivalent to N3, but significantly superior to other N levels. Conversely, in DSR with RDN (120 kg N ha⁻¹), the cost–benefit ratio surpassed that of TR by 16.96%, signifying DSR’s adaptability for more profitable rice cultivation in the region. This research provides valuable insights into optimizing nitrogen management practices for TR and DSR, thereby enhancing rice crop performance and economic returns.
... Furthermore, Baresel et al. [92] used a digital consumer camera to detect wheat leaf chlorophyll content and compared it with SPAD meter and reflectance spectrometry findings in three field experiments. The image analysis method comprises the segmentation and subsequent analysis of the foreground color, that is, only green plant parts; thus, analysis of canopies with a small degree of ground cover is also possible. ...
Article
Salinity is a key factor limiting agricultural production worldwide. Recent advances in field phenotyping have enabled the recording of the environmental history and dynamic response of plants by considering both genotype × environment (G×E) interactions and envirotyping. However, only a few studies have focused on plant salt tolerance phenotyping. Therefore, we analyzed the potential opportunities and major challenges in improving plant salt tolerance using advanced field phenotyping technologies. RGB imaging and spectral and thermal sensors are the most useful and important sensing techniques for assessing key morphological and physiological traits of plant salt tolerance. However, field phenotyping faces challenges owing to its practical applications and high costs, limiting its use in early generation breeding and in developing countries.
... Close relationships have been reported between leaf chlorophyll and nitrogen concentrations in a number of crops such as barley [5], rice [6], and wheat [7]. Traditional measurements of chlorophyll concentration have involved spectrophotometric measures of chlorophyll in acetone [8] or ethanol [9]. ...
Article
Full-text available
Image technologies have been used for real-time estimation of nitrogen (N) and leaf chlorophyll (Chl) concentrations as well as for photosynthetic properties. The aim of this work was to establish correlations between RGB values and chlorophyll and nitrogen concentrations in three ornamental potted plants. We evaluated the RGB values, nitrogen status, and chlorophyll concentrations in the leaves of Peperomia obtusifolia, Maytenus senegalensis, and Rosmarinus officinalis. The correlation between the RGB values and the chlorophyll and nitrogen concentrations in the leaves was different for each species, since baby rubber correlated with the R and G values, the confetti tree correlated with the G and B values, and rosemary correlated with the R, G, and B values. The correlation between the normalized RGB (rgb) values and the color parameters and the chlorophyll and nitrogen concentrations showed R2 values lower than 0.70 in all species. Moreover, the estimation of vegetation indices was not effective due to the lack of correlations between these indices and the chlorophyll and nitrogen concentrations in the leaves of each species. According to the findings, rosemary exhibited the best association between the RGB values and chlorophyll and nitrogen concentrations in the leaves.
... Those application rely on the principle that relationship among the spectral refl ectance of the leaves or canopy may be correlated with canopy growth traits, plant nutrients status and productivity, and the relationships can be described by exploring information contained in images (WANG et al., 2014). Many research efforts have been dedicated to the image processing software development capable of determining the plant nutrient status through images acquired by scanners, commercial digital cameras, or even by cellphone and smartphone cameras, generating information to support decision making on fertilization strategies (BARESEL et al., 2017;HU et al., 2013;LEE;LEE, 2013;MOHAN;GUPTA, 2019;ZILBERMAN et al., 2018). ...
Article
Full-text available
ABSTRACT The occurrence of water sources with high concentrations of salts is a common problem in the semi-arid region of north-eastern Brazil. The search for management strategies that can minimize the effect of salt stress on crops is therefore extremely important. As such, this study aimed to evaluate gas exchange and production in the yellow passion fruit ‘BRS GA1’as a function of irrigation strategies using brackish water and doses of potassium. The research was carried out under field conditions in São Domingos, in the state of Paraíba, Brazil, using a randomized block design in a 6 × 2 factorial scheme, with treatments comprising six irrigation strategies using water (irrigation with low-salinity water throughout the cycle – WS; irrigation with high-salinity water during the vegetative stage – VE; during the flowering stage – FL; the fruiting stage – FR; during successive vegetative/flowering stages - VE/FL; successive vegetative/fruiting stages - VE/FR) and two doses of potassium (60% and 100% of the recommended dose of 345 g K2O per plant per year), with four replications and three plants per plot. Two levels of water salinity (1.3 and 4.0 dS m−1) were used during different phenological stages of the crop. Irrigation with water at 4.0 dS m−1 reduced the leaf water potential, leaf osmotic potential, stomatal conductance, transpiration, and rate of CO2 assimilation of the yellow passion fruit, regardless of the irrigation strategy. The continuous salt stress during the vegetative and flowering stages compromised production in the yellow passion fruit.low-salinity water throughout the cycle – WS; irrigation with high-salinity water during the vegetative stage – VE; during the flowering stage – FL; the fruiting stage – FR; during successive vegetative/flowering stages - VE/FL; successive vegetative/fruiting stages - VE/FR) and two doses of potassium (60% and 100% of the recommended dose of 345 g K2O per plant per year), with four replications and three plants per plot. Two levels of water salinity (1.3 and 4.0 dS m−1) were used during different phenological stages of the crop. Irrigation with water at 4.0 dS m−1 reduced the leaf water potential, leaf osmotic potential, stomatal conductance, transpiration, and rate of CO2 assimilation of the yellow passion fruit, regardless of the irrigation strategy. The continuous salt stress during the vegetative and flowering stages compromised production in the yellow passion fruit.
... Those application rely on the principle that relationship among the spectral refl ectance of the leaves or canopy may be correlated with canopy growth traits, plant nutrients status and productivity, and the relationships can be described by exploring information contained in images (WANG et al., 2014). Many research efforts have been dedicated to the image processing software development capable of determining the plant nutrient status through images acquired by scanners, commercial digital cameras, or even by cellphone and smartphone cameras, generating information to support decision making on fertilization strategies (BARESEL et al., 2017;HU et al., 2013;LEE;LEE, 2013;MOHAN;GUPTA, 2019;ZILBERMAN et al., 2018). ...
Article
Full-text available
This study is based on the principle that vegetation indexes (VIs), derived from the RGB color model obtained from digital images, can be used to characterize spectral signatures and classify Brachiaria brizantha cv. Xaraés according to nitrogen status (N). From colorimetric data obtained from leaf blade images acquired in the fi eld, three artifi cial neural networks were evaluated according to the performance in the classifi cation of N status: Feedforward Backpropagation (FFBP), Cascade Forward Backpropagation (CFBP) and Radial Base function (RBFNN). Four N fertilization rates were applied to generate contrasting N contents in the plants. The youngest completely expanded leaves from 60 tillers were detached at each regrowth cycle of 28 days, thus their images and leaf N content were obtained. Samples were then classifi ed as defi cient (< 17 g N kg-1 leaf dry matter (DM), moderately defi cient (from 17.1 to 20.0 g N kg-1 DM), and suffi cient (> 20.1 g N kg-1 DM). The VIs were selected by principal component analysis and the performance of the networks evaluated by the accuracy. The accuracy in classifi cation obtained by the networks were 88%, 86% and 79% for FFBP, CFBP and RBFNN, respectively, indicating that the spectral signatures can be determined from images acquired in the fi eld. So, the proposed method could be used to develop a software that aims to monitor the status of N in real time, providing a fast and inexpensive tool for defi ning the time and the amount of N fertilizer, according to the pasture demand.
... G, B ? G, R ? G ? B (Hu et al. 2013), R/B, and G/B (Baresel et al. 2017) have also been shown to correlate well with Chl content. However, theoretical explanations for all such correlations from the physiological and photochemistry perspectives may not be possible. ...
Article
Chlorophyll (Chl) concentration is a reliable indicator of leaf nitrogen content and plant health status. Currently available methods for image-based Chl estimation require complex mathematical derivations and high-throughput imaging set-up along with multiplex image-preprocessing steps. Further, the influence of carotenoid (CAR)concentration has been largely ignored in the process. The present study describes a smartphone-based leaf image analysis method for real-time estimation of Chl concentration and Chl/CAR ratio. Color features were obtained from RGB (red, green, blue) images of spinach leaves using a smartphone, and inverse R and G values were calculated. Correlation analysis of color indices and photosynthetic pigment (PP) concentrations was performed, followed by principal component analysis (PCA). Linear mathematical modeling was performed for describing regression equations for predicting PP concentration. 1/R and 1/G indices showed strong positive linear correlation (0.93 < r2 < 0.96) with Chl and CAR concentrations, respectively. Furthermore, 1/R + 1/G and [1/R]/[1/G] presented strong positive linear correlation with Chl + CAR (r2 = 0.95) and Chl/CAR (r2 = 0.88), respectively. PCA confirmed the association of color indices with the respective PP features, which were subsequently estimated using the correlation models. A smartphone-based companion application was developed using the linear models for non-invasive, real-time estimation of Chl concentration and Chl/CAR ratio. 1/R and 1/G indices indicate the concentrations of Chl and CAR, respectively, via linear models. The smartphone application developed using the linear models enables real-time estimation of Chl concentration and Chl/CAR ratio without complicated image preprocessing steps or mathematical derivations.
... A number of researchers evaluated the potential of RGB and multispectral cameras to measure N status and they revealed that both imageries followed by reliable process can estimate the N status of the crop. 23,[31][32][33][34][35][36][37][38] The use of drones for crop condition monitoring in India is still in a nascent stage. This is primarily because very few studies have been conducted to develop remote sensing-based models for different crop conditions for different Indian crops and their varieties. ...
Article
Nitrogen is one of the essential nutrients required for crop growth, and hence should be applied efficiently for attaining optimum yield. To fulfil nitrogen need, absorbed nitrogen in the plant is required to be estimated. Various methods are available to estimate crop nitrogen such as tissue analysis using the methods of Kjeldahl and Dumas, which are accurate, but timeconsuming and destructive. Satellite imagery provides a more extensive field view. However, they are limited to their spatial and temporal resolution. Unmanned aerial vehicle (UAV) is emerging as a promising tool that can provide the status of crop nitrogen rapidly with high spatial and temporal resolution. The objective of the study was to evaluate UAV-based imageries to show nitrogen status and predict rice yield at different growth stages. The experiments were conducted using two rice cultivars, six nitrogen applications, two water management practices, and with three replications. Soil plant analysis development (SPAD) meter readings were collected at various growth stages. First, aerial imageries of experimental site were collected using an octocopter UAV equipped with a multispectral sensor that provides reflectance values in four different bands (red, green, red edge, and near-infrared) along with SPAD values for respective flight. Second, aerial images were processed in pix4D software, to identify the most appropriate vegetation index that shows nitrogen status variation in the field and to predict yield using different vegetation indices. Nine vegetation indices were considered: ratio vegetation index, normalized difference vegetation index, normalized green red difference index, red edge difference vegetation index, green ratio vegetation index, green normalized difference vegetation index (GNDVI), wide dynamic range vegetation index, transformed normalized difference vegetation index (TNDVI), and normalized difference red edge. After that, a linear regression model was developed between the representative index and SPAD values. Finally, linear regression models developed by using VI and SPAD values were evaluated and results revealed that GNDVI-based model simulates SPAD values with R 2 of 0.49, 0.49, and 0.74 at panicle, milky, and booting stages, respectively. It is also found that TNDVI-based linear regression model predicts yield with R 2 of 0.71 at milky stage.
... Chlorophyll content changes can be used to determine maximal photosynthetic capability, productivity, leaf growth level, and stress (Zhang et al. 2019). Furthermore, because much of the nitrogen in leaves is absorbed in chlorophyll, chlorophyll concentration provides an indirect estimate of nutrition status (Baresel et al. 2017). A growing number of abiotic and biotic disturbances are deteriorating and weakening forests throughout the world (Macpherson et al. 2017;Senf et al. 2017;Wilkaniec et al. 2021). ...
Article
Full-text available
Canopy Chlorophyll Content (CCC) is the most important biophysical parameter in a forest eco-system, since it determines plant production, stress, and photosynthetic capability. Plant adaptation monitoring in a changing environment necessitates analyzing long-term changes in CCC. Present study computes CCC using the PROSAIL Radiative Transfer Model Inversion (RTM) with Artificial Neural Networking (ANN). The data from the Sentinel 2A satellite was utilized for this purpose. The various vegetation indices (VIs) were derived from Landsat and Sentinel data. The Infrared Percentage vegetation index (IPVI) from Landsat 8 demonstrated a strong connection with CC (R2 = 0.8). Green vegetation index (GVI), Normalized difference index (NDI), and Pigment specific simple ratio index (PSSRa) exhibited good correlations with CCC. From 1997 to 2017, the correlation of IPVI with CCC was utilized to model the spatio-temporal variation of CCC. The negative trend and decrease of CCC was detected at a rate of − 1.2 g cm−2 year−1 throughout this 20-year period with 33% fall in chlorophyll concentration, indicating a substantial reduction in forest health. The primary differences observed in dense forest area CCC and change in agricultural CCC were minor. This dramatic drop in chlorophyll concentration creates a variety of photosynthetic vulnerabilities in the forest ecosystem, resulting in forest degradation that may have unintended consequences for humans and wildlife.
... Knowing plant N status is important for optimal N fertilizer management the N status of crops is measured with destructive plant samples obtained from the field. Although chemical analyses of plant N uptake are very accurate, it remains difficult to obtain this information over a large area and the amount of sampling and laboratory expenses are often prohibitive (Baresel et al., 2017;Frels et al., 2018). Therefore non-destructive spectral analyses have evolved for crop monitoring and the diagnosis of plant N status without intensive sampling (Verrelst et al., 2015;Barmeier and Schmidhalter, 2017;Elsayed et al., 2018;Prey and Schmidhalter, 2019). ...
Article
Nitrogen (N) fertilization management plays an important role in optimizing crop growth and yield. Concerns over environmental risk require a quick, accurate, non-destructive determination of the N status of crops. Hyperspectral remote sensing allows a timely monitoring of the in-season crop N status. Although many spectral indices for assessing the N status of crops have been proposed, it is still necessary to further optimize the central bands, since they often vary with plant cultivars and species. To improve this, we identified optimized three-band spectral indices for estimating the canopy N uptake of corn and wheat. Experiments were conducted from 2009 to 2011 by evaluating and testing optimized three-band spectral indices for estimating the N status of wheat cultivars grown in Germany and China and corn cultivars grown in China. The indices generally enabled more robust predictions compared to published indices. The central bands suitable for assessing the canopy N uptake were 768, 740, and 548 nm for corn; 876, 736 and 550 nm for wheat; and 846, 732 and 536 nm for corn and wheat combined. Both wheat and corn assessed individually as well as in combination where sharing similar wavebands reflected by the species-specific and interspecies-specific optimized three-band spectral indices, e.g. the wavelengths 740, 736 and 732 nm were identified as optimum for corn, wheat and their combination, respectively. The validation results suggest that predictions using the optimized three-band N planar domain index (NPDI) delivered the highest coefficient of determination (R 2 = 0.86) and the lowest root mean square error (RMSE, 20.1 kg N ha-1) and relative error (RE, 18.7 %). The optimized NPDI consistently estimated canopy N uptake of both corn and wheat alone and in combination. Therefore, the optimized three-band algorithm is an attractive tool for optimizing and identifying central bands. Our algorithm may allow the design of improved N diagnosis systems and enhance the application of ground-and satellite-based sensing.
Article
Non-destructive RGB sensor-based estimation of chlorophyll content has significant uses in crop demand-based nitrogen fertilizer application in precision agriculture and plant phenotyping for crop improvement. In this work, we estimated the chlorophyll content of wheat rapidly and no-destructively with high-throughput phenotyping by using RGB-based vegetation indices. The RGB images were captured using the automated phenotyping and imaging platform at Nanaji Deshmukh Plant Phenomics Centre, New Delhi, India. Then the RGB images were analysed and used to calculate 39 RGB vegetative indices reported in the literature. We examined the RGB vegetative indices from RGB images and 16 machine-learning models were used to measure total chlorophyll content and the models were evaluated by coefficient of determination (R2), correlation coefficient (r), root means square error (RMSE) to select the best estimation model. The results showed that several RGB indices and total chlorophyll content values showed a highly significant correlation, with a high correlation reaching 0.84. In the prediction model, the highest precision was obtained with the BRNN model (R2 = 0.71, r = 0.84, RMSE = 2.52). These indicated that the best model (BRNN) can precisely estimate whole plant total chlorophyll content in wheat from digital RGB images, which implies RGB indices have potential for low-cost wheat leaf chlorophyll estimation. Since RGB sensor is low-cost sensor, this model can be used drone-based and mobile based applications in precision agriculture and plant phenotyping. This non-invasive method can uncover potential genes for chlorophyll accumulation by estimating the pigment status of plants at various phenological stages.
Article
Accurate assessment of chlorophyll content in plants is crucial for understanding their health and physiological performance. In this study, we investigated the feasibility of using digital image analysis as a non-destructive approach to determine chlorophyll levels in Sonchus oleraceus L., a less explored species with potential to serve as an alternative commercial crop tolerant to abiotic stress, satisfying specific niche needs. We compared the effectiveness of digital image analysis with traditional biochemical methods considering seasonal variations and cultivation systems. Image analysis involved assessing leaf color intensities, focusing on the Green color scale. Results revealed a strong correlation between leaf color intensities in digital images and chlorophyll levels determined by biochemical methods. Additionally, it was observed that the Green color scale exhibited the highest intensity, emerging as the most promising for future analyses. Statistical analysis revealed that the Spearman correlation method was most suitable for this type of study, especially in the second growing season. In conclusion, digital image analysis proved to be an effective and non-destructive tool for assessing chlorophyll content in S. oleraceus L., with potential application in agricultural and environmental studies.
Article
Full-text available
Robotics and variable rate technology (VRT) have shown great potential for site-specific weed management (SSWM), but these technologies face several challenges like accurate weed identification, high initial cost, integration with existing farming practices, and the need for the development of precise and robust system for real-time application. Given the lack of comprehensive information on the current state of robotic technologies for weed management, this review study aims to provide an extensive overview of the individual components associated with SSWM technologies, such as autonomous navigation, imaging sensors, and weeding approaches. To perform this study, 91 technical papers were selected that involve the status, challenges, and potential improvements of individual components integrated to develop an efficient weed management technology. The key outcomes can be summarized as: (a) real-time kinematic-enabled global positioning system (RTK-GPS) and machine vision guidance systems are widely adopted technology for automatic row guidance because of their low cost, high accuracy, and easy accessibility, (b) GPS technology face challenges due to signal interference and dependency on satellite availability for accurate positioning, (c) sensor fusion has been suggested as a potential solution to improve the accuracy of positioning and weed identification, (d) deep learning techniques for weed identification offer significant promise in enhancing the feasibility of implementing robotic weed control in real-time, (e) due to robust nature of hyperspectral imaging, research in weed identification for real-time in open-field conditions is still lagging, (f) plethora of university-based research studies are limited to model optimization and simulation studies, and (g) integrated weed management practices, such as a combined application of mechanical and laser weeding, could be the future of robotic weed control. This review will serve as a valuable reference to extension educators, growers, researchers, and stakeholders for adopting state-of-the-art and sustainable practices in SSWM.
Article
Full-text available
Wetlands, as a crucial component of terrestrial ecosystems, play a significant role in global ecological services. Aboveground biomass (AGB) is a key indicator of the productivity and carbon sequestration potential of wetland ecosystems. The current research methods for remote-sensing estimation of biomass either rely on traditional vegetation indices or merely perform integer-order differential transformations on the spectra, failing to fully leverage the information complexity of hyperspectral data. To identify an effective method for estimating AGB of mixed-wetland-plant communities, we conducted field surveys of AGB from three typical wetlands within the Crested Ibis National Nature Reserve in Hanzhong, Shaanxi, and concurrently acquired canopy hyperspectral data with a portable spectrometer. The spectral features were transformed by applying fractional-order differentiation (0.0 to 2.0) to extract optimal feature combinations. AGB prediction models were built using three machine learning models, XGBoost, Random Forest (RF), and CatBoost, and the accuracy of each model was evaluated. The combination of fractional-order differentiation, vegetation indices, and feature importance effectively yielded the optimal feature combinations, and integrating vegetation indices with feature bands enhanced the predictive accuracy of the models. Among the three machine-learning models, the RF model achieved superior accuracy using the 0.8-order differential transformation of vegetation indices and feature bands (R² = 0.673, RMSE = 23.196, RPD = 1.736). The optimal RF model was visually interpreted using Shapley Additive Explanations, which revealed that the contribution of each feature varied across individual sample predictions. Our study provides methodological and technical support for remote-sensing monitoring of wetland AGB.
Article
Full-text available
As a vital pigment for photosynthesis in rice, chlorophyll content is closely correlated with growth status and photosynthetic capacity. The estimation of chlorophyll content allows for the monitoring of rice growth and facilitates precise management in the field, such as the application of fertilizers and irrigation. The advancement of hyperspectral remote sensing technology has made it possible to estimate chlorophyll content non-destructively, quickly, and effectively, offering technical support for managing and monitoring rice growth across wide areas. Although hyperspectral data have a fine spectral resolution, they also cause a large amount of information redundancy and noise. This study focuses on the issues of unstable input variables and the estimation model’s poor applicability to various periods when predicting rice chlorophyll content. By introducing the theory of harmonic analysis and the time-frequency conversion method, a deep neural network (DNN) model framework based on wavelet packet transform-first order differential-harmonic analysis (WPT-FD-HA) was proposed, which avoids the uncertainty in the calculation of spectral parameters. The accuracy of estimating rice chlorophyll content based on WPT-FD and WPT-FD-HA variables was compared at seedling, tillering, jointing, heading, grain filling, milk, and complete periods to evaluate the validity and generalizability of the suggested framework. The results demonstrated that all of the WPT-FD-HA models’ single-period validation accuracy had coefficients of determination (R2) values greater than 0.9 and RMSE values less than 1. The multi-period validation model had a root mean square error (RMSE) of 1.664 and an R2 of 0.971. Even with independent data splitting validation, the multi-period model accuracy can still achieve R2 = 0.95 and RMSE = 1.4. The WPT-FD-HA-based deep learning framework exhibited strong stability. The outcome of this study deserves to be used to monitor rice growth on a broad scale using hyperspectral data.
Article
Full-text available
The accurate instance segmentation of individual crop plants is crucial for achieving a high-throughput phenotypic analysis of seedlings and smart field management in agriculture. Current crop monitoring techniques employing remote sensing predominantly focus on population analysis, thereby lacking precise estimations for individual plants. This study concentrates on maize, a critical staple crop, and leverages multispectral remote sensing data sourced from unmanned aerial vehicles (UAVs). A large-scale SAM image segmentation model is employed to efficiently annotate maize plant instances, thereby constructing a dataset for maize seedling instance segmentation. The study evaluates the experimental accuracy of six instance segmentation algorithms: Mask R-CNN, Cascade Mask R-CNN, PointRend, YOLOv5, Mask Scoring R-CNN, and YOLOv8, employing various combinations of multispectral bands for a comparative analysis. The experimental findings indicate that the YOLOv8 model exhibits exceptional segmentation accuracy, notably in the NRG band, with bbox_mAP50 and segm_mAP50 accuracies reaching 95.2% and 94%, respectively, surpassing other models. Furthermore, YOLOv8 demonstrates robust performance in generalization experiments, indicating its adaptability across diverse environments and conditions. Additionally, this study simulates and analyzes the impact of different resolutions on the model’s segmentation accuracy. The findings reveal that the YOLOv8 model sustains high segmentation accuracy even at reduced resolutions (1.333 cm/px), meeting the phenotypic analysis and field management criteria.
Chapter
Crop diseases cause a decrease in the quantity and quality of agricultural goods. Current problem-solving methods have evolved dramatically courtesy of machine learning (ML). Predicting the outcome with a trustworthy accuracy score, significantly lowering crop loss. A classification is a process that groups leaves according to their various morphological characteristics. The process of choosing a classification method is always challenging because the quality of the results can differ depending on the input data. K-nearest neighbor classifier (KNN), probabilistic neural network (PNN), genetic algorithm, support vector machine (SVM), and principal component analysis, artificial neural network (ANN), and fuzzy logic are a few examples of diverse classification algorithms. Classifications of plant leaf diseases have many uses in a variety of industries, including agriculture and biological research. Switching from one disease control strategy to another presents significant challenges for farmers. The control of illnesses through effective management strategies can be facilitated by early information on crop health and disease identification. Rice crop productivity will increase, thanks to this method. The classic method of plant disease detection and identification in real life is naked eye observation. We will also focus on the use of several of these strategies on openly accessible datasets to comprehend how the methods for reducing dimensionality function to provide us precise findings for crop disease prediction and support their management.
Article
Çeltik, temel bir gıda kaynağıdır ve endüstride sıkça kullanılan nadir bitkilerden biridir. Çeltik yaprak hastalıklarının erken teşhisi, ekin hasarını en aza indirmek için büyük önem taşımaktadır. Son yıllarda, derin öğrenme tabanlı bilgisayar destekli sistemler, ziraat sektöründe oldukça önem kazanmış ve çeşitli uygulamalarda etkin rol almıştır. Bu sistemler, hastalıkların erken teşhis edilmesine yardımcı olmakla kalmayıp, aynı zamanda tarım alanında çalışanlara da ikincil bir yardımcı olarak katkı sağlamaktadır. Bu çalışma, çeltik yapraklarında bulunan hastalıkların erken teşhisinde derin öğrenme yöntemlerinin etkinliğini araştırmayı amaçlamaktadır. Bu amaç doğrultusunda, VGG, ResNet, DenseNet, EfficientNet, Inception ve Xception gibi en popüler evrişimsel sinir ağları (CNN), halka açık Paddy Doctor veri seti üzerinde değerlendirilmiştir. Her bir modele, veri ön işleme, veri artırma, hiper-parametre optimizasyonu ve öğrenme aktarımı gibi güncel teknikler uygulanarak test setindeki teşhis doğruluğunun başarımı arttırılmıştır. Ayrıca her bir mimarideki modellerin birbirine ve diğer mimarilerdeki modellere göre çeltik yapraklarındaki hastalıkların teşhisindeki başarımları detaylı bir şekilde karşılaştırılmıştır. Deneysel sonuçlar, EfficientNetv2_Small modelinin %98.01 test doğruluğu ve %97.99 F1-skor değerleriyle tüm modellerden daha iyi performans sergilediğini ve literatürdeki diğer çalışmaları geride bıraktığını göstermiştir. Bu çalışma, CNN mimarilerinin yüksek bir performans gösterdiğini ve bu tür hastalıkların erken teşhisinde ziraat mühendislerine ve çiftçilere etkili bir şekilde yardımcı olabileceğini göstermektedir
Article
Full-text available
The system fertilization approach emerged to improve nutrient use efficiency in croplands. This new fertilization concept aims at taking advantage of nutrient cycling within an agroecosystem to obtain maximum production from each nutrient unit. To monitor this eect, methodologies such as the Normalized Dierence Vegetation Index (NDVI) are promising to evaluate plant biomass and nutrient content. We evaluated the use of NDVI as a predictor of shoot biomass, P and K uptake, and yield in soybean. Treatments consisted of two production systems [integrated crop-livestock system (ICLS) and cropping system (CS)] and two periods of phosphorus (P) and potassium (K) fertilization (crop fertilization—P and K applied at soybean sowing—and system fertilization—P and K applied in the pasture establishment). NDVI was evaluated weekly from the growth stage V2 up to growth stage R8, using the Greenseeker® canopy sensor. At the growth stages V4, V6, R2, and R4, plants were sampled after NDVI evaluation for chemical analysis. Soybean yield and K uptake were similar between production systems and fertilization strategies (P > 0.05). Soybean shoot biomass and P uptake were, respectively, 25.3% and 29.7% higher in ICLS compared to CS (P < 0.05). For NDVI, an interaction between the production system and days after sowing (P < 0.05) was observed. NDVI increased to 0.95 at 96 days after sowing in CS and to 0.92 at 92 days after sowing in ICLS. A significant relationship between NDVI and shoot biomass, and P and K uptake was observed (P < 0.05). Our results show that the vegetation index NDVI can be used for estimating shoot biomass and P and K uptake in the early growth stages of soybean crops, providing farmers with a new tool for evaluating the spatial variability of soybean growth and nutrition.
Chapter
Plant breeding primarily focuses on improving agronomy traits, e.g. yield, quality, host plant resistance to pathogens and pests, and abiotic stress tolerance; however, the methods for their genetic improvement are being rapidly enhanced through genomics and phenomics. In the Genomics–Phenomics–Agronomy (G‐P‐A) paradigm, diverse research approaches have been conducted to bridge any two of these elements, and recently, all of them together. This review first highlights the progress to link (1) genomics to agronomy, (2) genomics to phenomics, and (3) phenomics to agronomy. Secondly, the G‐P‐A domain is dissected into different layers, each addressing the three elements simultaneously. These layers include genetic dissection through gene mapping using genome‐wide association studies and genomic selection using best linear unbiased prediction, Bayesian approaches, and machine learning. The objective of the review is to help readers to grasp the core developments among the exponentially growing literature in each of these fields. Through this review, the connections among the three elements of the G‐P‐A paradigm are coherently integrated toward the prospect of sustainable development of agronomy traits through both genomics and phenomics.
Article
Full-text available
Canopy covers can be measured using destructive (visual) and non-destructive methods (spectral indices, photogrammetry, visual assessment, and quantum sensor). The precision of crop cover estimation, however, is dependent on the selection of appropriate methods. Studies were conducted at the Indian Grassland and Fodder Research Institute, Jhansi to compare the forage crops canopy cover estimated using photogrammetry software (Canopeo and SamplePoint) and visual assessments. Assessments were performed in three summer crops (corn, cowpea, and sorghum), two winter crops (Egyptian clover, and oats), and bare ground condition. For each plot, three nadir images (directly above the canopy) were captured using digital cameras from a height of 1.5 m above the soil surface between 10 AM to 2 PM on bright sunny days. The results indicated that the relationships between visual assessment and Canopeo (regression coefficient, (R² = 0.96), visual assessment and SamplePoint (0.96), and Canopeo and SamplePoint (0.98) were linear when data were pooled across all the crops. SamplePoint and Canopeo is further, appropriate for cowpea (Pearson coefficient (R = 0.99 and 0.94), oats (0.92 and 0.97), and sorghum (0.46 and 0.51), respectively. SamplePoint and Canopeo are not suitable for berseem (-0.15) and corn (-0.61), respectively, due to dead residues after the first harvest in berseem and taller corn might have influenced the image quality. Therefore, the stage of the crop, the height of the crop, and dead residues around the plants can greatly influence the estimation of crop cover. In conclusion, the results indicated that this photogrammetry software can be used for non-destructive crop canopy measurement with the above-mentioned precautions in the forage crops tested. • Forage canopy cover is estimated generally by visual scoring, and the outcome varies widely from person to person. • Photogrammetry methods (Canopeo and SamplePoint) were positiviely correlated with visual scoring for cowpea, oats, and sorghum. • However, Canopeo and SamplePoint may not suitable for taller crops like corn and ratoon crops like berseem.
Article
In agriculture, identifying the deficient or excess nutrients in the leaf is a significant concern for attaining the higher yield and productivity of any crop. Several image processing techniques have been developed by various researchers for detecting and reducing the deficiency in plant leaves. Some of those detection mechanisms are not scalable, expensive, and may lead to an irreversible stage of a crop while they fail to detect the nutrient level of a leaf at an earlier period. When complexity increases during the prediction process, it leads to untimely predictions that will affect the quality and production of a crop. The Nutrient Range Analysis Based on Greenness (NRAG) system is proposed to detect and classify nitrogen nutrient of groundnut leaves. Median filter is selected for its highest peak signal to noise ratio (PSNR) value to reduce the noise occurrence from the collected groundnut dataset during pre-processing. Greenness percentage of groundnut leaves are calculated with colour features and found that it is directly proportional to nitrogen (N) nutrient. Then those features are given to the developed Convolution Neural Network (CNN) based Holding Vector Network (HVN) model, which helps to detect and classify the three different ranges of nutrients, including low N nutrient, no N deficiency, and excess N nutrient. The obtained results shows that the developed CNN-based HVN model achieved 95% training accuracy and 92% validation accuracy. The developed model can be adopted by all types of farmers in real time to predict nitrogen range because of its high accuracy, scalability, and cost effectiveness.
Article
Full-text available
The paddy crop is the most essential and consumable agricultural produce. Leaf disease impacts the quality and productivity of paddy crops. Therefore, tackling this issue as early as possible is mandatory to reduce its impact. Consequently, in recent years, deep learning methods have been essential in identifying and classifying leaf disease. Deep learning is used to observe patterns in disease in crop leaves. For instance, organizing a crop’s leaf according to its shape, size, and color is significant. To facilitate farmers, this study proposed a Convolutional Neural Networks-based Deep Learning (CNN-based DL) architecture, including transfer learning (TL) for agricultural research. In this study, different TL architectures, viz. InceptionV3, VGG16, ResNet, SqueezeNet, and VGG19, were considered to carry out disease detection in paddy plants. The approach started with preprocessing the leaf image; afterward, semantic segmentation was used to extract a region of interest. Consequently, TL architectures were tuned with segmented images. Finally, the extra, fully connected layers of the Deep Neural Network (DNN) are used to classify and identify leaf disease. The proposed model was concerned with the biotic diseases of paddy leaves due to fungi and bacteria. The proposed model showed an accuracy rate of 96.4%, better than state-of-the-art models with different variants of TL architectures. After analysis of the outcomes, the study concluded that the anticipated model outperforms other existing models.
Article
Discrimination of tea plant (Camellia sinensis L.) varieties is of significant value for the efficient management of cultivation and resources optimization of tea industry. Phenotypic and spectral characteristics of tea plants are important indicators for determining the quality of tea cultivars to a certain extent. However, few studies have used spectral image information to identify tea plant varieties. The aim of this research was the discrimination of 16 types of high-yield tea plant varieties using a multispectral camera. This methodology involved image registration, calibration, segmentation, information extraction, and data fusion. The hue (H), saturation (S), and value (V), texture information, and several spectral vegetation indices were acquired from the multispectral image of the tea plant canopy. The successive projection algorithm (SPA) was used to analyze the original parameters. Three classification methods were applied to tea plant variety discrimination: Bayes discriminant analysis (BDA), support vector machine (SVM), and extreme learning machine (ELM). The results indicated that SPA based on fusing data combined with the SVM classification model, achieved a feasible method to identify tea plant varieties. Additionally, the method achieved accuracy in the training, test, and validation sets, reaching 97.00%, 90.52%, and 88.67%, respectively. This study proposed a new perspective on multispectral image information as an identifier of tea plant varieties. The explored model will be helpful for the development of portable instruments for commercial applications in variety identification and phenotype recognition of tea plants.
Article
Full-text available
The accurate and rapid estimation of the aboveground biomass (AGB) of rice is crucial to food security. Unmanned aerial vehicles (UAVs) mounted with hyperspectral sensors can obtain images of high spectral and spatial resolution in a quick and effective manner. Integrating UAV-based spatial and spectral information has substantial potential for improving crop AGB estimation. Hyperspectral remote-sensing data with more continuous reflectance information on ground objects provide more possibilities for band selection. The use of band selection for the spectral vegetation index (VI) has been discussed in many studies, but few studies have paid attention to the band selection of texture features in rice AGB estimation. In this study, UAV-based hyperspectral images of four rice varieties in five nitrogen treatments (N0, N1, N2, N3, and N4) were obtained. First, multiple spectral bands were used to identify the optimal bands of the spectral vegetation indices, as well as the texture features; next, the vegetation index model (VI model), the vegetation index combined with the corresponding-band textures model (VI+CBT model), and the vegetation index combined with the full-band textures model (VI+FBT model) were established to compare their respective rice AGB estimation abilities. The results showed that the optimal bands of the spectral and textural information for AGB monitoring were inconsistent. The red-edge and near-infrared bands demonstrated a strong correlation with the rice AGB in the spectral dimension, while the green and red bands exhibited a high correlation with the rice AGB in the spatial dimension. The ranking of the monitoring accuracies of the three models, from highest to lowest, was: the VI+FBT model, then the VI+CBT model, and then the VI model. Compared with the VI model, the R2 of the VI+FBT model and the VI+CBT model increased by 1.319% and 9.763%, respectively. The RMSE decreased by 2.070% and 16.718%, respectively, while the rRMSE decreased by 2.166% and 16.606%, respectively. The results indicated that the integration of vegetation indices and textures can significantly improve the accuracy of rice AGB estimation. The full-band textures contained richer information that was highly related to rice AGB. The VI model at the tillering stage presented the greatest sensitivity to the integration of textures, and the models in the N3 treatment (1.5 times the normal nitrogen level) gave the best AGB estimation compared with the other nitrogen treatments. This research proposes a reliable modeling framework for monitoring rice AGB and provides scientific support for rice-field management.
Article
Full-text available
RZ (2022). Effects of nitrogen application rates on the spatio-temporal variation of leaf SPAD readings on the maize canopy. The Journal of Agricultural Science 1-13. https://doi. Abstract The spatio-temporal variation of leaf chlorophyll content is an important crop phenotypic trait that is of great significance for evaluating crop productivity. This study used a soil-plant analysis development (SPAD) chlorophyll meter for non-destructive monitoring of leaf chlorophyll dynamics to characterize the patterns of spatio-temporal variation in the nutritional status of maize (Zea mays L.) leaves under three nitrogen treatments in two culti-vars. The results showed that nitrogen levels could affect the maximum leaf SPAD reading (SPAD max) and the duration of high SPAD reading. A rational model was used to measure the changes in SPAD readings over time in single leaves. This model was suitable for predicting the dynamics of the nutrient status for each leaf position under different nitrogen treatments , and model parameter values were position dependent. SPAD max at each leaf decreased with the reduction of nitrogen supply. Leaves at different positions in both cultivars responded differently to higher nitrogen rates. Lower leaves (8th-10th positions) were more sensitive than the other leaves in response to nitrogen. Monitoring the SPAD reading dynamic of lower leaves could accurately characterize and assess the nitrogen supply in plants. The lower leaves in nitrogen-deficient plants had a shorter duration of high SPAD readings compared to nitrogen-sufficient plants; this physiological mechanism should be studied further. In summary, the spatio-temporal variation of plant nitrogen status in maize was analysed to determine critical leaf positions for potentially assisting in the identification of appropriate agronomic management practices, such as the adjustment of nitrogen rates in late fertilization.
Article
The color of plant leaves can be assessed qualitatively by color charts or after processing of digital images. This pilot study employed a novel pocket-sized sensor to obtain the color of plant leaves. In order to assess its performance, a color-dependent parameter (SPAD index) was used as the dependent variable, since there is a strong correlation between SPAD index and greenness of plant leaves. A total of 1,872 fresh and intact leaves from 13 crops were analyzed using a SPAD-502 m and scanned using the Nix™ Pro color sensor. The color was assessed via RGB and CIELab systems. The full dataset was divided into calibration (70% of data) and validation (30% of data). For each crop and color pattern, multiple linear regression (MLR) analysis and multivariate modeling [least absolute shrinkage and selection operator (LASSO), and elastic net (ENET) regression] were employed and compared. The obtained MLR equations and multivariate models were then tested using the validation dataset based on r, R², root mean squared error (RMSE), and mean absolute error (MAE). In both RGB and CIELab color systems, the Nix™ Pro color sensor was able to differentiate crops, and the SPAD indices were successfully predicted, mainly for mango, quinoa, peach, pear, and rice crops. Validation results indicated that ENET performed best in most crops (e.g., coffee, corn, mango, pear, rice, and soy) and very close to MLR in bean, grape, peach, and quinoa. The correlation between SPAD and greenness is crop-dependent. Overall, the Nix™ Pro color sensor was a fast, sensible and an easy way to obtain leaf color directly in the field, constituting a reliable alternative to digital camera imagery and associated image processing.
Article
Computer Vision Systems (CVS) offer a non-destructive and contactless tool to assign visual quality level to fruit and vegetables and to estimate some of their internal characteristics. The innovative CVS described in this paper exploits the combination of image processing techniques and machine learning models (Random Forests) to assess the visual quality and predict the internal traits on unpackaged and packaged rocket leaves. Its performance did not depend on the cultivation system (traditional soil or soilless). The same CVS, exploiting its machine learning components, was able to build effective models for either the classification problem (visual quality level assignment) and the regression problems (estimation of senescence indicators such as chlorophyll and ammonia contents) just by changing the training data. The experiments showed a negligible performance loss on packaged products (Pearson’s linear correlation coefficient of 0.84 for chlorophyll and 0.91 for ammonia) with respect to unpackaged ones (0.86 for chlorophyll and 0.92 for ammonia). Thus, the non-destructive and contactless CVS represents a valid alternative to destructive, expensive and time-consuming analyses in the lab and can be effectively and extensively used along the whole supply chain, even on packaged products that cannot be analyzed using traditional tools.
Article
Full-text available
The relationship between early ground cover and grain yield in winter wheat has not yet been fully understood. In a winter wheat breeding program, selection for early ground cover is traditionally made using visual scoring. Although visual scoring is preferred as a phenotypic screening tool by wheat breeders, its output may not be reliable due to requiring experience. The smartphone camera-based digital image technique can be recommended as a feasible, reliable, repeatable, affordable, and speed selection tool for early ground cover in wheat as an alternative to visual scoring. For this purpose, two wheat trials were conducted in the 2017-2018 and 2019-2020 seasons. In both seasons, 215 wheat genotypes in total, together with three checks from spring wheat, were tested under the rain-fed conditions of the spring wheat zone of Turkey. All wheat genotypes tested were grouped into spring, facultative, and winter growth types (habit) using visual scoring. Simultaneously, photos were taken from each plot with a smartphone camera. Then, the early ground covers (%) were estimated using the smartphone camera-based digital image technique, from which the relationships between grain yield, visual scoring, and early ground cover were estimated. In both seasons, significant negative relationships between grain yield and visual scoring (r=−0.679** and r=−0.704**, respectively), significant positive relationships between grain yield and early ground cover (r=0.745** and r=0.747**, respectively), significant negative relationships between visual scoring and early ground cover (r=−0.862** and r=−0.926**, respectively) were detected. Also, the broad sense heritability values for variables measured in both seasons (0.51 and 0.85 for early ground cover, 0.91 and 0.94 for visual scoring, and 0.86 and 0.69 for grain yield, respectively) were calculated. In this study, we found that both genotypic variation (i.e., the different growth types) within winter wheat genotypes and the selection environment (i.e., testing winter wheat genotypes in spring wheat zone rather than in winter wheat zone) had causal effects on the emergence of a positive relationship between early ground cover and grain yield. As a result, thanks to this study, it has been proved that the smartphone-based digital image technique can be used as a selection tool for early ground cover in winter wheat.
Article
Aims Flower ultraviolet (UV) reflectance strongly influences insects’ ability to detect flowers and locate pollen/nectar sources. Although included in the TRY database, the lack of a cost‐effective standardised method of measurement hampers the availability of information on this key floral trait. Digital photography and image processing allow for a novel approach to flower UV reflectance measurement that is both accessible and accurate. Location European semi‐natural grasslands. Methods We used a UV‐LED light and a mirrorless camera with a filter, which represents a low‐cost equipment for narrow emission/reflection photography (350–380 nm). Flowers were photographed with two standards of known reflection, and UV reflectance values were obtained using open‐source image processing software. We measured UV reflectance for 57 plant species typical of European semi‐natural grasslands. Results Our values substantially matched the categorical classes obtained by analogue photography available in TRY and showed a highly significant relationship with spectrophotometric measures. Conclusions The method proposed here overcomes the one based on analogue photography and subjective visual estimates, and represents an easy and low‐cost alternative to spectrophotometry. It may promote the standardised measurement of flower UV reflectance and broaden the information of this trait globally, meeting the needs of functional ecology and trait‐based community assembly studies.
Article
Full-text available
Citrus Huanglongbing (HLB), also named citrus greening disease, occurs worldwide and is known as a citrus cancer without an effective treatment. The symptoms of HLB are similar to those of nutritional deficiency or other disease. The methods based on single-source information, such as RGB images or hyperspectral data, are not able to achieve great detection performance. In this study, a multi-modal feature fusion network, combining a RGB image network and hyperspectral band extraction network, was proposed to recognize HLB from four categories (HLB, suspected HLB, Zn-deficient, and healthy). Three contributions including a dimension-reduction scheme for hyperspectral data based on a soft attention mechanism, a feature fusion proposal based on a bilinear fusion method, and auxiliary classifiers to extract more useful information are introduced in this manuscript. The multi-modal feature fusion network can effectively classify the above four types of citrus leaves and is better than single-modal classifiers. In experiments, the highest accuracy of multi-modal network recognition was 97.89% when the amount of data was not very abundant (1,325 images of the four aforementioned types and 1,325 pieces of hyperspectral data), while the single-modal network with RGB images only achieved 87.98% recognition and the single-modal network using hyperspectral information only 89%. Results show that the proposed multi-modal network implementing the concept of multi-source information fusion provides a better way to detect citrus HLB and citrus deficiency.
Article
With the development of imaging devices and image processing algorithms, numerous features have come to be used for the estimation of total nitrogen content (TNC) in plants. However, higher-dimensional inputs contain more correlated variables that can detrimentally affect model performance. In this study, a hybrid feature selection approach was developed for TNC estimation in Aquilaria sinensis. A low-cost modified digital camera with external filters was used to capture canopy images. Three feature selection methods, namely, random forest (RF), Pearson correlation coefficient (PCC)-based feature selection, and sequential backward selection (SBS), were combined into two hybrid feature selection algorithms (RF_SBS and PCC_SBS). In addition, three regression algorithms were used in hybrid feature selection process: random forest regression (RFR), support vector regression (SVR), and partial least squares regression (PLSR). The hybrid feature selection process consists of two steps. First, the lowest number of dimensions is sought based on the feature ranking. Then, SBS is used to find the best feature combinations. Compared with the original models, the R² values of the RF-SBS-based models are improved by 0.094 (RF_SBS_RFR), 0.190 (RF_SBS_SVR), and 0.116 (RF_SBS_PLSR), while the R² values of the PCC-SBS-based models are improved by 0.055 (PCC_SBS_RFR), 0.092 (PCC_SBS_SVR) and 0.128 (PCC_SBS_PLSR). Finally, the two best TNC estimation models are found to be PCC_SBS_PLSR, with an R² of 0.863, and RF_SBS_SVR, with an R² of 0.872. The proposed hybrid feature selection approach not only has great capacity to improve estimation accuracy but also can reduce model complexity by choosing the best feature subset.
Article
Full-text available
In precision agriculture, the nitrogen level is significantly important for establishing phenotype, quality and yield of crops. It cannot be achieved in the future without appropriate nitrogen fertilizer application. Moreover, a convenient and real-time advance technology for nitrogen nutrition diagnosis of crops is a prerequisite for an efficient and reasonable nitrogen-fertilizer management system. With the development of research on plant phenotype and artificial intelligence technology in agriculture, deep learning has demonstrated a great potential in agriculture for recognizing nondestructive nitrogen nutrition diagnosis in plants by automation and high throughput at a low cost. To build a nitrogen nutrient-diagnosis model, muskmelons were cultivated under different nitrogen levels in a greenhouse. The digital images of canopy leaves and the environmental factors (light and temperature) during the growth period of muskmelons were tracked and analyzed. The nitrogen concentrations of the plants were measured, we successfully constructed and trained machine-learning- and deep-learning models based on the traditional backpropagation neural network (BPNN), the emerging convolution neural network (CNN), the deep convolution neural network (DCNN) and the long short-term memory (LSTM) for the nitrogen nutrition diagnosis of muskmelon. The adjusted determination coefficient (R2) and mean square error (MSE) between the predicted values and measured values of nitrogen concentration were adopted to evaluate the models’ accuracy. The values were R2 = 0.567 and MSE = 0.429 for BPNN model; R2 = 0.376 and MSE = 0.628 for CNN model; R2 = 0.686 and MSE = 0.355 for deep convolution neural network (DCNN) model; and R2 = 0.904 and MSE = 0.123 for the hybrid model DCNN–LSTM. Therefore, DCNN–LSTM shows the highest accuracy in predicting the nitrogen content of muskmelon. Our findings highlight a base for achieving a convenient, precise and intelligent diagnosis of nitrogen nutrition in muskmelon.
Article
Full-text available
Intermediate wheatgrass (IWG) [Thinopyrum intermedium (Host) Barkworth & Dewey] is the first commercially produced perennial grain crop in the United States. Intercropping legumes with IWG has the potential to enhance dual‐purpose grain and forage production and contribute to weed control in organic management systems. We compared IWG with annual winter wheat (Triticum aestivum L.) in monoculture and intercropped with red clover (Trifolium pratense L.) in a 3‐yr experiment in central New York. Grain yield of IWG was lower than wheat in all years, partly due to lower tiller fertility and seed size in IWG. Compared with grain yield of 1,212 kg ha–1 in the first year, IWG grain yield was 83% (202 kg ha–1) and 64% lower (441 kg ha–1) in the second and third years, respectively. Intermediate wheatgrass straw production increased 40% from 5,541 to 7,785 kg ha–1 over 3 yr while wheat straw yield declined from 5,167 to 3,533 kg ha–1. Red clover did not affect grain or straw yield of either crop but reduced weed biomass and weed species richness. Weed communities in IWG plots were dominated by perennial grasses by the second year of production, whereas annual weeds were dominant in wheat throughout the experiment. Preventing establishment of perennial weeds that will persist in perennial grain cropping systems should be a management priority. High forage production observed when comparing IWG and wheat suggest opportunities for including IWG in integrated crop–livestock systems where IWG's higher forage yield and quality has higher utility.
Article
Full-text available
Chlorophyll content is a widely used parameter for nutritional status diagnosis in sugarcane. This study aimed to develop a predictive model of chlorophyll content in sugarcane seedlings using spectral imagery analysis within the electromagnetic spectrum visible range. The experiment was carried out in a split-plot design, with two fertilization rates and three sugarcane cultivars. For chlorophyll analysis, 144 leaves were collected from seedlings. Chlorophyll contents were extracted and measured by SPAD-502 meter. Spectral images within the range of 480 to 710 nm were analyzed using reflectance, absorbance (white source), and fluorescence (source at 405 and 470 nm) responses. Predictive models were developed using multivariate regression methods such as Principal Component Regression and Partial Least Squares Regression. We chose the best model through absorbance response using variable selection and the PLSR method (R2P = 0.718 and RMSEP = 7.665). The wavelengths of 480, 490, 500, 600, 630, and 640 nm were identified as the best for total chlorophyll content determination. The spectral image processing-based method can provide a chlorophyll measurement equivalent to SPAD, with the advantage of having a higher spatial coverage over the entire leaf area. Besides, it can also support automation of the chlorophyll measurement in greenhouses. non-destructive methods; chemometrics; principal component regression; partial least squares regression
Article
Full-text available
Weeds are unwanted plants that can reduce crop yields by competing for water, nutrients, light, space, and carbon dioxide, which need to be controlled to meet future food production requirements. The integration of drones, artificial intelligence, and various sensors, which include hyperspectral, multi-spectral, and RGB (red-green-blue), ensure the possibility of a better outcome in managing weed problems. Most of the major or minor challenges caused by weed infestation can be faced by implementing remote sensing systems in various agricultural tasks. It is a multi-disciplinary science that includes spectroscopy, optics, computer, photography, satellite launching, electronics, communication, and several other fields. Future challenges, including food security, sustainability, supply and demand, climate change, and herbicide resistance, can also be overcome by those technologies based on machine learning approaches. This review provides an overview of the potential and practical use of unmanned aerial vehicle and remote sensing techniques in weed management practices and discusses how they overcome future challenges.
Article
Full-text available
The determination of leaf nitrogen content (LNC) by indirect methods is essential for silvicultural management of forest crops. The application of photography or rapid measurement equipment, such as chlorophyll index (soil-plant analysis development-SPAD), is increasingly used due to its low-cost, ease of estimation and accuracy. Therefore, the aim of this study was to estimate foliar nitrogen content from nondestructive methods in plantations of Eucalyptus tereticornis and Eucalyptus saligna using three urea treatments (120 kg N ha-1 , 240 kg N ha-1 and a control treatment without urea). For each treatment, 10 trees were selected, including four for the validation of the equations. The LNC was directly evaluated for color with the CIEL*a*b* model, photographic measurement with the RGB model, SPAD measurement and destructive estimation of nitrogen in leaves. The results showed negative relationships with the L* (luminosity) and b* (trend from yellow to green) indices, while the a* (red to green trend) index was discarded, with SPAD positive relationships were found with LNC and RGB space. In the R and B indices, the greatest negative relationships were found. It was determined that the multivariate equation Y=a+b 1 x 1 +b 2 x 2 +…+b n x n can be used for this type of study. It was also determined that the LNC=0.389+0.026SPAD model was the optimum for E. tereticornis and the LNC=3.826-0.001R-0.10B equation was the optimum for E. saligna.
Article
Full-text available
Rice disease has serious negative effects on crop yield, and the correct diagnosis of rice diseases is the key to avoid these effects. However, the existing disease diagnosis methods for rice are neither accurate nor efficient, and special equipment is often required. In this study, an automatic diagnosis method was developed and implemented in a smartphone app. The method was developed using deep learning based on a large dataset that contained 33,026 images of six types of rice diseases: leaf blast, false smut, neck blast, sheath blight, bacterial stripe disease, and brown spot. The core of the method was the Ensemble Model in which submodels were integrated. Finally, the Ensemble Model was validated using a separate set of images. Results showed that the three best submodels were DenseNet-121, SE-ResNet-50, and ResNeSt-50, in terms of several attributes, such as, learning rate, precision, recall, and disease recognition accuracy. Therefore, these three submodels were selected and integrated in the Ensemble Model. The Ensemble Model minimized confusion among the different types of disease, reducing misdiagnosis of the disease. Using the Ensemble Model to diagnose six types of rice diseases, an overall accuracy of 91% was achieved, which is considered to be reasonably good, considering the appearance similarities in some types of rice disease. The smartphone app allowed the client to use the Ensemble Model on the web server through a network, which was convenient and efficient for the field diagnosis of rice leaf blast, false smut, neck blast, sheath blight, bacterial stripe disease, and brown spot.
Article
Full-text available
Estimation of plant canopy using low-altitude imagery can help monitor the normal growth status of crops and is highly beneficial for various digital farming applications such as precision crop protection. However, extracting 3D canopy information from raw images requires studying the effect of sensor viewing angle by taking into accounts the limitations of the mobile platform routes inside the field. The main objective of this research was to estimate wheat (Triticum aestivum L.) leaf parameters, including leaf length and width, from the 3D model representation of the plants. For this purpose, experiments with different camera viewing angles were conducted to find the optimum setup of a mono-camera system that would result in the best 3D point clouds. The angle-control analytical study was conducted on a four-row wheat plot with a row spacing of 0.17 m and with two seeding densities and growth stages as factors. Nadir and six oblique view image datasets were acquired from the plot with 88% overlapping and were then reconstructed to point clouds using Structure from Motion (SfM) and Multi-View Stereo (MVS) methods. Point clouds were first categorized into three classes as wheat canopy, soil background, and experimental plot. The wheat canopy class was then used to extract leaf parameters, which were then compared with those values from manual measurements. The comparison between results showed that (i) multiple-view dataset provided the best estimation for leaf length and leaf width, (ii) among the single-view dataset, canopy , and leaf parameters were best modeled with angles vertically at −45° and horizontally at 0° (VA −45, HA 0), while (iii) in nadir view, fewer underlying 3D points were obtained with a missing leaf rate of 70%. It was concluded that oblique imagery is a promising approach to effectively estimate wheat canopy 3D representation with SfM-MVS using a single camera platform for crop monitoring. This study contributes to the improvement of the proximal sensing platform for crop health assessment.
Article
Chlorophyll (Chl) concentration is a reliable indicator of leaf nitrogen content and plant health status. Currently available methods for image-based Chl estimation require complex mathematical derivations and high-throughput imaging set-up along with multiplex image-preprocessing steps. Further, the influence of carotenoid (CAR)concentration has been largely ignored in the process. The present study describes a smartphone-based leaf image analysis method for real-time estimation of Chl concentration and Chl/CAR ratio. Color features were obtained from RGB (red, green, blue) images of spinach leaves using a smartphone, and inverse R and G values were calculated. Correlation analysis of color indices and photosynthetic pigment (PP) concentrations was performed, followed by principal component analysis (PCA). Linear mathematical modeling was performed for describing regression equations for predicting PP concentration. 1/R and 1/G indices showed strong positive linear correlation (0.93 < r2 < 0.96) with Chl and CAR concentrations, respectively. Furthermore, 1/R + 1/G and [1/R]/[1/G] presented strong positive linear correlation with Chl + CAR (r2 = 0.95) and Chl/CAR (r2 = 0.88), respectively. PCA confirmed the association of color indices with the respective PP features, which were subsequently estimated using the correlation models. A smartphone-based companion application was developed using the linear models for non-invasive, real-time estimation of Chl concentration and Chl/CAR ratio. 1/R and 1/G indices indicate the concentrations of Chl and CAR, respectively, via linear models. The smartphone application developed using the linear models enables real-time estimation of Chl concentration and Chl/CAR ratio without complicated image preprocessing steps or mathematical derivations.
Article
Full-text available
Advances in proximal hyperspectral sensing tools, chemometric techniques, and data-driven modeling have enhanced precision irrigation management by facilitating the monitoring of several plant traits. This study investigated the performance of remote sensing indices derived from thermal and red-green-blue (RGB) images combined with stepwise multiple linear regression (SMLR) and an integrated adaptive neuro-fuzzy inference system with a genetic algorithm (ANFIS-GA) for monitoring the biomass fresh weight (BFW), biomass dry weight (BDW), biomass water content (BWC), and total tuber yield (TTY) of two potato varieties under 100%, 75%, and 50% of the estimated crop evapotranspiration (ETc). Results showed that the plant traits and indices varied significantly between the three irrigation regimes. Furthermore, all of the indices exhibited strong relationships with BFW, CWC, and TTY (R2 = 0.80–0.92) and moderate to weak relationships with BDW (R2 = 0.25–0.65) when considered for each variety across the irrigation regimes, for each season across the varieties and irrigation regimes, and across all data combined, but none of the indices successfully assessed any of the plant traits when considered for each irrigation regime across the two varieties. The SMLR and ANFIS-GA models gave the best predictions for the four plant traits in the calibration and testing stages, with the exception of the SMLR testing model for BDW. Thus, the use of thermal and RGB imaging indices with ANFIS-GA models could be a practical tool for managing the growth and production of potato crops under deficit irrigation regimes.
Article
Full-text available
Accurate estimation of crop nitrogen status is essential for effective management of nitrogen applications in precision agriculture. Excessive nitrogen application pollutes underground and running water what leads to serious environmental problems. Variable rate technology for fertilizer application has been developed to reduce environmental risks and increase fertilizer use efficiency. For precise application of this technology, it requires online determination of plant nutrient status in the field. In this study, for the estimation of sugar beet leaf nitrogen, status color image analysis was performed for the first time. The experiment was carried out in a phytotron and nitrogen was applied at six levels to the sugar beet grown in pots. Leaf nitrogen concentration was measured by a SPAD-502 chlorophyll meter. For nitrogen status estimation, a neural-network model and linear regression model were developed based on red (R), green (G) and blue (B) components of the color image captured with a conventional digital camera. The obtained results demonstrated that the neural network model is capable of sugar beet leaf nitrogen estimation with reasonable accuracy. The coefficient of determination (R2) and mean square error (MSE) between the estimated and the measured SPAD values, which were obtained from validation tests, were 0.94 and 0.006, respectively. The results also indicated that the neural network model was performed with higher accuracy than the linear regression model, with R 2= 0.88.
Article
Full-text available
In contrast to high-throughput genotyping which can manage a large number of plants at relatively low cost, phenotyping of many individual genotypes in field trials is still laborious and expensive. Early plant vigour, as an early selection criterion, is a trait that is visually scored due to a lack of suitable phenotyping methods for an accurate detection of this trait in large field trials. A high-throughput phenotyping technique for scoring early plant vigour would enhance the breeding process. This study was conducted to develop a method for scoring phenotypic differences in early plant vigour of 50 winter wheat (Triticum aestivum L.) cultivars in a 2-years experiment using a vehicle based multispectral active sensor and two commercially available active sensors, GreenSeeker and CropCircle. Pixel analysis of RGB images revealed to be the most feasible and superior method compared to other possible reference methods. A comparison between the two years 2011 and 2012 confirmed that early plant vigour was affected by genotypic differences. A novel spectral plant vigour index (EPVI) was found to accurately reflect the plant vigour at tillering. Different methods were applied to identify optimal combinations of wavelengths to predict early plant vigour, including multivariate modelling and prediction, contour maps for identifying all possible simple ratios and testing of combined indices. The EPVI and the relative amount of green pixels (RAGP) derived from digital images were significantly related with r2 = 0.98 to each other in both years. A total of 200 plots, 12 m in length, could be measured within 75 min. The EPVI was shown to be an accurate scoring method for the high-throughput screening of large field trials. The rapidity and accuracy of this novel method may contribute to enhanced selection at early growth stages.
Article
Full-text available
Yield and grain protein concentration (GPC) represent crucial factors in the global agricultural wheat (Triticum aestivum L.) production and are predominantly determined via carbon and nitrogen metabolism, respectively. The maintenance of green leaf area and the onset of senescence (Osen) are expected to be involved in both C and N accumulation and their translocation into grains. The aim of this study was to identify stay-green and early senescence phenotypes in a field experiment of 50 certified winter wheat cultivars and to investigate the relationships among Osen, yield and GPC. Colour measurements on flag leaves were conducted to determine Osen for 20 cultivars and partial least square regression models were used to calculate Osen for the remaining 30 cultivars based on passive spectral reflectance measurements as a high-throughput phenotyping technique for all varieties. Using this method, stay-green and early senescence phenotypes could be clearly differentiated. A significant negative relationship between Osen and grain yield (r² = 0.81) was observed. By contrast, GPC showed a significant positive relationship to Osen (r² = 0.48). In conclusion, the high-throughput character of our proposed phenotyping method should help improve the detection of such traits in large field trials as well as help us reach a better understanding of the consequences of the timing of senescence on yield.
Article
Full-text available
Measuring the nitrogen nutrition status of plants is useful for nitrogen fertiliser management. As nitrogen is one of the main structural components of chlorophyll, its nutrition status is highly correlated with the greenness of leaves. This paper proposes and evaluates a new low-cost method to estimate the N-nutrition status of plants using digital colour image analysis. A method has been developed in which principal component analysis is applied to digital images to calculate a greenness index using RGB components of the colour image, which yields an estimate of the amount of N in the plant. To evaluate its quality, we calculated the correlation between the index and measurements obtained with a SPAD-502 chlorophyll meter, normally used in decision-making in fertiliser management. The performance of the proposed index is better than that of others previously investigated. Furthermore, the capacity of our index to predict N deficiencies affecting barley yield was equal to or better than that of SPAD measurements under our experimental conditions.
Article
Full-text available
Spectral measurements can be used to detect the canopy N status and allow locally adjusted N fertilizer application during the growing season. The present study investigates the potential of a tractor-based field spectrometer with a quadrilateral view optical setup to detect aboveground biomass dry weight and total aerial N in winter wheat (Triticum aestivum L.) in a 3-yr field experiment. The measuring time and the weather conditions were chosen, to be able to compare instrument performance under various sky conditions (sunny, cloudy, hazy) and at different solar zenith angles. New and known vegetations indices were tested in dense, high-yielding crop stand (average yields 6-10 Mg ha(-1)) receiving 0 to 210 kg N ha(-1) to detect aboveground biomass dry weight and total aerial N. Validations were performed on large calibration areas of 25 m(2) matching exactly the field of view from the sensor. The results obtained show that strong relationships (R(2) = 0.90) exist between reflectance indices and total aerial N from the end frittering at growth stage CS 29 to flowering at CS 71, whereas aboveground biomass dry weight could be detected with a R(2) of 0.75. A new simple index R(780)/R(740) was developed with similar potential to the red edge inflection point (REIP). The optical setup allows for measurements to be highly independent of time of day and sky conditions. Consequently the system promises to contribute to improved N management in heterogeneous fields.
Article
Full-text available
Several sensor systems are available for ground-based remote sensing in crop management. Vegetation indices of multiple active and passive sensors have seldom been compared in determining plant health. This work describes a study comparing active and passive sensing systems in terms of their ability to recognize agronomic parameters. One bi-directional passive radiometer (BDR) and three active sensors, including the Crop Circle, GreenSeeker, and an active flash sensor (AFS), were tested for their ability to assess six destructively determined crop parameters. Over 2 years, seven wheat (Triticum aestivum L.) cultivars were grown with nitrogen supplies varying from 0 to 220 kg ha−1. At three developmental stages, the crop reflectance was recorded and sensor-specific indices were calculated and related to N levels and the crop parameters, fresh weight, dry weight, dry matter content, as percent of dry weight to fresh weight, N content, aboveground N uptake, and the nitrogen nutrition index. The majority of the tested indices, based on different combinations of wavelengths in the visible and near infrared spectral ranges, showed high r2-values when correlated with the crop parameters. However, the accuracy of discriminating the influence of varying N levels on various crop parameters differed between sensors and showed an interaction with growing seasons and developmental stage. Visible- and red light-based indices, such as the NDVI, simple ratio (R780/R670), and related indices tended to saturate with increasing crop stand density due to a decreased sensitivity of the spectral signal. Among the destructively assessed biomass parameters, the best relationships were found for N-related parameters, with r2-values of up to 0.96. The near infrared-based index R760/R730 was the most powerful and temporarily stable index indicating the N status of wheat. This index was delivered by the BDR, Crop Circle, and AFS. Active spectral remote sensing is more flexible in terms of timeliness and illumination conditions, but to date, it is bound to a limited number of indices. At present, the broad spectral information from bi-directional passive sensors offers enhanced options for the future development of crop- or cultivar-specific algorithms.Highlights• Evaluation of multiple sensors’ predictive performance and capability in discriminating effects of four N rates on crop health of seven wheat cultivars grown over two seasons. • The accuracy of discriminating the influence of varying N levels on the nitrogen status and crop parameters differed between sensors and showed interactions with years and developmental stages. • Visible and red-light based indices showed a decreased sensitivity of the spectral signals at increasing crop biomass density. • NIR/NIR based indices of all sensors were most powerful and temporarily stable expressing the N status.
Article
Full-text available
The ability to assess green biomass is of particular interest in a number of wheat breeding environments. However, the measurement of this and similar traits is either tedious and time-consuming or requires the use of expensive, sophisticated equipment, such as field-based spectroradiometers to measure vegetation indices (VIs). Here, conventional digital cameras are proposed as affordable and easy-to-use tools for gathering field data in wheat breeding programmes. Using appropriate software, a large set of images can be automatically processed to calculate a number of VIs, based on the performance of simple colour operations on each picture. The purpose of this study was to identify a set of picture-derived vegetation indices (picVIs) and to evaluate their performance in durum wheat trials growing under rainfed and supplementary irrigation conditions. Here, zenithal pictures of each plot were obtained roughly 2 weeks after anthesis, and the picVIs that were calculated were compared with the normalised difference vegetation index (NDVI), an index derived from spectroradiometrical measurements, and with the grain yield (GY) from the same plots. The picVIs that performed best were the Hue, CIE-Lab a* and CIE-Luv u* components of the average colour of each picture, the relative green area (GA) and the ‘greener area’, similar to GA but excluding the more yellowish-green pixels. Our results showed a high correlation between all these picVIs and the NDVI. Moreover, in rainfed conditions, each picVI provided an estimation of GY similar to or slightly better than that provided by the NDVI. However, in irrigated conditions during anthesis, neither these picVIs nor the NDVI provided a good estimation of GY, apparently because of the saturation of the VI response in conditions of complete soil cover and high plant density.
Article
Full-text available
Crop management can be optimized and nitrogen (N) losses can be reduced with a better knowledge of soil-nitrogen availability, especially if this information becomes directly available on-site in a fast and cost-effective way. In this paper, simple on-farm methods to determine nitrate-N in field-moist soil samples immediately after sampling are described. The procedures include volumetric soil sampling, extraction based on manual shaking with tap water as universally available extractant, filtering soil/extractant mixtures on-site, on-site determination of the soil water content, and reflectometric nitrate measurements based on test strips. Using correction factors can compensate the impact of the temperature during the final nitrate measurement. An excellent agreement was found between the developed quick-test procedures and the standard laboratory procedure. The proposed quick-test has great potential to enable economical savings for farmers as well as benefiting the environment.
Article
Full-text available
Host plant resistance is an important component to the management of potato late blight,Phytophthora infestans (Mont.) de Bary. Assessment of potato lines (Solanum tuberosum L.) with various levels of resistance toP. infestans (US8, A2 genotype) were evaluated in field trials, greenhouse controlled environment chambers and inoculated tuber reactions. Five lines (AWN86514-2, B0692-4, B0718-3, Jacqueline Lee, and B0288-17) with strong foliar resistance to late blight were identified in these inoculated field trials. Greenhouse controlled environment chamber studies allowed resistant and susceptible lines to be distinguished, but the 1998 greenhouse results did not correlate well with field data. Four lines (A084275-3, Bzura, MSG007-1, and MSG297-4RD) evaluated by a digital image analysis technique demonstrated tuber resistance based upon average reflective index values in the inoculated tuber studies. Tuber resistance did not correlate with field foliar resistance. Based upon these results, field assessment of foliar reaction toP. infestans provides the best measure for assessing late blight resistance in potato. Tuber resistance to late blight can be identified among lines with varying levels of foliar resistance.
Article
Full-text available
A low-altitude platform utilising a 1.8-m diameter tethered helium balloon was used to position a multispectral sensor, consisting of two digital cameras, above a fertiliser trial plot where wheat (Triticum spp.) was being grown. Located in Cecil Plains, Queensland, Australia, the plot was a long-term fertiliser trial being conducted by a fertiliser company to monitor the response of crops to various levels of nutrition. The different levels of nutrition were achieved by varying nitrogen application rates between 0 and 120 units of N at 40 unit increments. Each plot had received the same application rate for 10 years. Colour and near-infrared images were acquired that captured the whole 2 ha plot. These images were examined and relationships sought between the captured digital information and the crop parameters imaged at anthesis and the at-harvest quality and quantity parameters. The statistical analysis techniques used were correlation analysis, discriminant analysis and partial least squares regression. A high correlation was found between the image and yield (R2 = 0.91) and a moderate correlation between the image and grain protein content (R2 = 0.66). The utility of the system could be extended by choosing a more mobile platform. This would increase the potential for the system to be used to diagnose the causes of the variability and allow remediation, and/or to segregate the crop at harvest to meet certain quality parameters.
Article
Full-text available
We conduct an exhaustive survey of image thresholding methods, categorize them, express their formulas under a uniform notation, and finally carry their performance comparison. The thresholding methods are categorized according to the information they are exploiting, such as histogram shape, measurement space clustering, entropy, object attributes, spatial correlation, and local gray-level surface. 40 selected thresholding methods from various categories are compared in the context of nondestructive testing applications as well as for document images. The comparison is based on the combined performance measures. We identify the thresholding algorithms that perform uniformly better over nonde- structive testing and document image applications. © 2004 SPIE and IS&T. (DOI: 10.1117/1.1631316)
Article
Full-text available
Reliable, precise and accurate estimates of disease severity are important for predicting yield loss, monitoring and forecasting epidemics, for assessing crop germplasm for disease resistance, and for understanding fundamental biological processes including co-evolution. Disease assessments that are inaccurate and/or imprecise might lead to faulty conclusions being drawn from the data, which in turn can lead to incorrect actions being taken in disease management decisions. Plant disease can be quantified in several different ways. This review considers plant disease severity assessment at the scale of individual plant parts or plants, and describes our current understanding of the sources and causes of assessment error, a better understanding of which is required before improvements can be targeted. The review also considers how these can be identified using various statistical tools. Indeed, great strides have been made in the last thirty years in identifying the sources of assessment error inherent to visual rating, and this review highlights ways that assessment errors can be reduced—particularly by training raters or using assessment aids. Lesion number in relation to area infected is known to influence accuracy and precision of visual estimates—the greater the number of lesions for a given area infected results in more overestimation. Furthermore, there is a widespread tendency to overestimate disease severity at low severities (
Article
Full-text available
Measuring percent occurrence of objects from digital images can save time and expense relative to conventional field measurements. However, the accuracy of image analysis had, until now, not reached the level of the best conventional field measurements. Additionally, most image-analysis software programs require advanced user training to successfully analyze images. Here we present a new software program, 'SamplePoint,' that provides the user a single-pixel sample point and the ability to view and identify the pixel context. We found SamplePoint to allow accuracy comparable with the most accurate field-methods for ground-cover measurements. Expert use of the program requires minimal training and its ease of use allows rapid measurements from image data. We recommend SamplePoint for calibrating the threshold-detection level of image-analysis software or for making direct measurements of percent occurrence from digital images.
Article
Full-text available
Relationships between chlorophyll concentration ([chl]) and SPAD values were determined for birch, wheat, and potato. For all three species, the relationships were non-linear with an increasing slope with increasing SPAD. The relationships for birch and wheat were strong (r 2 ∼ 0.9), while the potato relationship was comparatively weak (r 2 ∼ 0.5). Birch and wheat had very similar relationships when the chlorophyll concentration was expressed per unit leaf area, but diverged when it was expressed per unit fresh weight. Furthermore, wheat showed similar SPAD–[chl] relationships for two different cultivars and during two different growing seasons. The curvilinear shape of the SPAD–[chl] relationships agreed well with the simulated effects of non-uniform chlorophyll distribution across the leaf surface and multiple scattering, causing deviations from linearity in the high and low SPAD range, respectively. The effect of non-uniformly distributed chlorophyll is likely to be more important in explaining the non-linearity in the empirical relationships, since the effect of scattering was predicted to be comparatively weak. The simulations were based on the algorithm for the calculation of SPAD-502 output values. We suggest that SPAD calibration curves should generally be parameterised as non-linear equations, and we hope that the relationships between [chl] and SPAD and the simulations of the present study can facilitate the interpretation of chlorophyll meter calibrations in relation to optical properties of leaves in future studies.
Conference Paper
Full-text available
In agronomy, image processing techniques are more and more used to detect crop, weeds, diseases ... We proposed to study the feasibility to use color and/or texture analysis to evaluate the number of wheat ears per m² to simplify the manual countings currently done. In this paper we present firstly the use of color and texture image processing together to detect the ears, before to propose and compare different texture image segmentation techniques based on feature extraction by first and higher order statistical methods. The extracted features are used for unsupervised pixel classification to obtain the different classes in the image, before to use the k-means algorithm. Three methods have been tested with very heterogeneous results, except the run length technique for which the results are close to the manual countings (66% error). The hypothesis took into account for the textural analysis methods are currently modify to justify them more accurately, especially concerning the number of classes and the size of the analysis window.
Article
For the image acquisition of fungal diseases on cereal leaves and for the detection of weeds in crops a multi spectral camera system was developed. The system consists of a filter wheel in front of a CCD-camera. On the filter wheel four filters with different transmittance were mounted. Using an electric motor the filters were positioned directly in front of the camera lens. The digital images which were obtained in this way were combined by an image processing software using division and subtraction. In the resulting images leaf diseases could be distinguished exactly. A segmentation of the plants for weed recognition was even possible with varying illumination conditions of image acquisition although an unchanged image processing algorithm was used. Furthermore the concept of a camera for the online acquisition of bispectral images for the detection of weeds is presented.
Article
Shape feature analyses were performed on binary images originally obtained from color images of 10 common weeds, along with corn and soybeans, found in the Midwest. Features studied were roundness, aspect, perimeter/thickness, elongatedness, and seven invariant central moments (ICM), for each plant type and age up to 45 days after emergence. Shape features were generally independent of plant size, image rotation, and plant location within most images. The ability to discriminate between monocots and dicots was most evident between 14 and 23 days using these features. Shape features that best distinguished these plants were aspect and first invariant central moment (ICM1), which classified 60 to 90% of the dicots from the monocots. Using Analysis of Variance and Tukey's multiple comparison tests, shape features did not change significantly for most species over the study period. This information could be very useful in the future design of advanced spot spraying applications.
Article
Improvement of nitrogen (N) use efficiency is urgently needed since excessive application of N fertilizer has been widespread in small-scale fields in China, causing great losses of N fertilizer and environmental pollution. In the present study, a simple technology, termed the Green Window Approach (GWA), to optimize N strategies for cereal crops is presented. The GWA represents an on-field demonstration site visualizing the effects of incremental N levels and enables farmers to conduct such a trial within their own fields. The lowest N rate that achieves no visible change in plant growth or biomass shows the optimal N requirement of crops. Therefore the objective was to develop the key procedures of GWA and to evaluate the effects of its application in cereal crops on grain yield, N use efficiency and economic benefit. A total of seven GWA trials were performed from 2009 to 2011 on farmers’ irrigated wheat fields in the North China Plain. The GWA consisted of eight small plots placed in a compact layout on a well-accessible part of the field. Plot size varied from 2·5×2·5 to 4×4 m2, depending on the size and shape of each field. All GWA plots received basal nitrogen (N), phosphorus (P) and potassium (K) rates of 30 kg N/ha (except for the nil-N plot), 80 kg P2O5/ha and 100 kg K2O/ha. Nitrogen supplies, including residual soil nitrate in 0–90 cm determined at Zadoks growth stages (GS) 21–23 in early spring and the split-topdressing N at GS 21–23 and GS 41–52, were incrementally increased from 0 to 420 kg N/ha. The remaining part of the field still received farmers’ customary fertilization (FCF). Optimal N rate could be estimated as the lowest N rate that achieved no visible change in plant growth at GS 60–73. Compared with FCF area, grain yield was increased by 13% to a maximum or near maximum value of 5·8 t/ha, optimal N rate was sharply decreased by 69% to 116 kg N/ha, apparent N recovery was greatly increased from 11 to 46%, whereas the cost of fertilizer input was decreased by 57% to 1045 Chinese Renminbi (RMB)/ha (162 US/ha),theprofitofgrainyieldwasincreasedby13/ha), the profit of grain yield was increased by 13% to 12 211 RMB/ha (1891 US/ha) and the net economic benefits were increased by 60% to 7473 RMB/ha (1157 US$/ha). Most importantly, the GWA does not need laboratory facilities, complicated procedures or professional knowledge of N balances, and farmers can easily understand and use GWA by themselves.
Article
Timely and accurate quantification of aerial nitrogen (N) uptake in crops is important for the calculation of regional N balances and the study of the N budget in agro-ecosystems. Experiments in the present study were conducted from 2007 to 2011 to remotely estimate the aerial N uptake of diverse winter wheat cultivars grown in contrasting climatic and geographic zones in China and Germany. Potentials and limitations of hyperspectral indices obtained from (i) optimized algorithms and (ii) 15 representative indices reported in the literature were tested for stability in estimating the aerial N uptake of winter wheat across different growth stages, cultivars, sites and years. Growth stage, cultivar, N application rates, site and year greatly influenced the relationship between hyperspectral indices and aerial N uptake. The optimized hyperspectral indices generally had more robust aerial N uptake prediction abilities than the published indices. Compared with the algorithms of all possible two-band combinations and red-edge position-based algorithms, area-based algorithms for a three-band optimized combination were more stable in deriving the aerial N uptake of winter wheat. Optimized algorithms can potentially be implemented in future aerial N uptake monitoring by hyperspectral sensing.
Article
Many spectral indices have been proposed to derive plant nitrogen (N) nutrient indicators based on different algorithms. However, the relationships between selected spectral indices and the canopy N content of crops are often inconsistent. The goals of this study were to test the performance of spectral indices and partial least square regression (PLSR) and to compare their use for predicting canopy N content of winter wheat. The study was conducted in cool and wet southeastern Germany and the hot and dry North China Plain for three winter wheat growing seasons. The canopy N content of winter wheat varied from 0.54% to 5.55% in German cultivars and from 0.57% to 4.84% in Chinese cultivars across growth stages and years. The best performing spectral indices and their band combinations varied across growth stages, cultivars, sites and years. Compared with the best performing spectral indices, the average value of the R2 for the PLSR models increased by 76.8% and 75.5% in the calibration and validation datasets, respectively. The results indicate that PLSR is a potentially useful approach to derive canopy N content of winter wheat across growth stages, cultivars, sites and years under field conditions when a broad set of canopy reflectance data are included in the calibration models. PLSR will be useful for real-time estimation of N status of winter wheat in the fields and for guiding farmers in the accurate application of their N fertilisation strategies.
Article
Leaf photosynthesis and rhizobial nitrogen fixation are the two metabolic processes of utmost importance to legume growth and development. As these processes are closely related to each other, measuring of leaf chlorophyll content can provide information on the nodulation and nitrogen fixation status of crop plants. In the present investigation, a number of soybean breeding lines consisting of near-isogenic families which are genetically segregating for the nodulation trait were utilized in field experiments carried out across three growing seasons at Vienna, Austria. For phenotyping leaf chlorophyll content, the Minolta SPAD spectrometer was applied in parallel to a simple leaf digital image analysis procedure based on a commercial digital still camera. The main objectives of the research included the comparison of SPAD metering and image analysis for determination of chlorophyll content, phenotyping of the soybean nodulation vs. non-nodulation characteristic with respect to leaf, agronomic and seed traits, and relating both chlorophyll and image analysis data to seed quality characteristics. Nodulating and non-nodulating soybean lines significantly differed in chlorophyll content from the V5 (five leaves fully developed) soybean developmental stage onwards. Apart from chlorophyll content, leaf size, plant height, number of pods per plant, 1000-seed weight, and seed protein and oil content were also affected by nodulation type. The chlorophyll content of soybean leaves as determined by SPAD metering was significantly correlated (r=−0.937) to the green color value (RGB color model) of leaf image analysis at the R3 (beginning of pod growth) soybean developmental stage. Both chlorophyll content and leaf image analysis parameters were correlated to 1000-seed weight, seed protein and seed oil content. Thus, it appears that these leaf parameters related to photosynthesis and nitrogen fixation could be utilized to determine the nitrogen status of a soybean crop and subsequently in forecasting seed quality parameters of the harvest product.
Article
Digital image analysis is an objective and nondestructive method potentially capable of providing accurate and precise estimates of disease resistance components. This study was conducted to quantify components of partial resistance to crown rust through the analysis of sequential digital images of inoculated leaves of adult oat plants, and to compare components found in two sources of resistance unrelated genetically. Uredinium density, relative infection frequency, latent period, days to first pustule appearance, uredinium size, and disease progress rates were assessed on three oat lines (RS-line 3W-UR-9-3b, MN-841801, Starter) in two greenhouse experiments. Resistant lines had fewer and smaller uredinia, and these appeared later than in the susceptible check. Relative infection frequency, latent period, and uredinium size were equally important components in the expression of the partial resistance to crown rust, and the two sources of resistance could not be differentiated by any of the variables studied. The analysis of sequential digital images of diseased leaves produced precise estimates of partial resistance components and disease progress rates.
Article
Digital image analysis could be a rapid and precise technique for estimating legume proportions in grass swards. In 2004, we conducted a pot study to evaluate a digital image analysis (DIA) system for estimation of legume dry matter (DM) contribution in legume-grass mixtures. Examination of pure swards and binary legume-grass mixtures of red clover (Trifolium pratense L.), white clover (T repens L.), alfalfa (Medicago sativa L.), and perennial ryegrass (Lolium perenne L.) took place after 35, 49, and 63 d of growth. To estimate the cover percentage of legumes in the swards, a total of 64 digital pictures were taken. The DM contribution of legumes (% of total biomass) showed a significant relationship with the proportion of image area covered by legumes (% of total area), which was classified visually. A DIA system for grayscale images was developed with the software Optimas. We found that DIA could be used to accurately predict legume contribution in mature swards. Legume contribution, as estimated by DIA, was significantly correlated with DM contribution of red clover (R(2) = 0.87), white clover (R(2) = 0.85), and alfalfa (R(2) = 0.79). Bare ground reduced the predictive ability of DIA in young or open swards. Use of DIA may be limited until we refine the method to deal with bare ground and different leaf shapes associated with various weed species.
Article
We have developed an image analysis system for mapping white clover pastures. The information from digital colour photographs is processed by software (Trifolium.exe) specially designed for the purpose. The software estimates the coverage of clover, weeds and bare ground, and the unidentified remainder of the total area is regarded as covered by grass. To evaluate the reliability of the estimates of clover content, the clover on paper printouts of non-processed images were marked manually and analysed by a photo scanner and commercially available software. The outputs from Trifolium.exe and the estimates from scanned manual markings were highly correlated (r 2=0.81). A sensitivity test was conducted to quantify the impact of changes of six user-adjustable parameters of Trifolium.exe. The software output of clover coverage was sensitive for changes in three, soil coverage for changes in one, and weed coverage for changes in all parameters. The fact that the digital image acquisition and analysis produce nearly continuous and exactly positioned data, implies further that it is a very appropriate tool for analyses of spatial dynamics in grass-clover pastures.
Article
Clover biomass and its percentage are both important measurements for performance trials of clover cultivars in binary mixtures with grass species. In the present study we developed multiple regression equations for estimating clover biomass and its percentage composition in binary mixtures of white clover (Trifolium repens L., three cultivars and one breeding line) and timothy (Phleum pratense L.). Plant height, cover, cover × plant height and dry matter ratio of white clover and cutting time were considered as potential explanatory variables for the regression equation. White clover cover was measured using a grid-point plate overlaid on a digital image of vegetation photographed from the top. The technique is considered to be useful for estimating herbage biomass and its percentage of white clover in a mixture with the grass because of high accuracy and precision in estimation and reduced time and labor requirements.
Article
Digital image processing has the potential to support the identification of plant species required for site-specific weed control in grassland swards. The present study focuses on the identification of one of the most invasive and persistent weed species on European grassland, the broad-leaved dock (Rumex obtusifolius L., R.o.), in complex mixtures of perennial ryegrass with R.o. and other herbs.A total of 108 digital photographs were obtained from a field experiment under constant recording geometry and illumination conditions. An object-oriented image classification was performed. Image segmentation was done by transforming the red, green, blue (RGB) colour images to greyscale intensity images. Based on that, local homogeneity images were calculated and a homogeneity threshold (0.97) was applied to derive binary images. Finally, morphological opening was performed. The remaining contiguous regions were considered to be objects. Features describing shape, colour and texture were calculated for each of these objects. A Maximum-likelihood classification was done to discriminate between the weed species. In addition, rank analysis was used to test how combinations of features influenced the classification result.The detection rate of R.o. varied with the training dataset used for classification. Average R.o. detection rates ranged from 71 to 95% for the 108 images, which included more than 3,600 objects. Misclassifications of R.o. occurred mainly with Plantago major (P.m.). Between 9 and 16% R.o. objects were classified incorrectly as P.m. and 17–24% P.m. objects were misclassified as R.o. The classification result was influenced by the defined object classes (R.o., P.m., T.o., soil, residue vs. R.o., residue). For instance, classification rates were 86–91% and 65–82% for R.o. exclusively and R.o. against the remaining herb species, respectively.
Article
Research into automatic image processing of digital plot photography has increased in recent years. However, in most studies only overall vegetation cover is estimated. In arid regions of the southwestern US, grass cover is typically a mixture of green and senescent plant material and it is important to be able to quantify both types of vegetation. Our objectives were to develop an image analysis approach for estimating fractional cover of green and senescent vegetation using very high-resolution ground photography, and to compare image-based estimates with line-point-intercept (LPI) measures. We acquired ground photography for 50 plots using an eight megapixel digital camera. The images were transformed from the RGB (red, green, blue) color space to the IHS (intensity, hue, saturation) color space. We used an object-based image analysis approach to classify the images into soil, shadow, green vegetation, and senescent vegetation. Shadow and soil were effectively masked out by using the intensity and saturation bands, and a nearest neighbor classification was used to separate green and senescent vegetation using intensity, hue and saturation as well as visible bands. Correlation coefficients between LPI- and image-based estimates for green and senescent vegetation were 0.88 and 0.95 respectively. Image analysis underestimated total and senescent vegetation by approximately 5%. The object-based image-processing approach is less labor and time intensive than the LPI method, is a viable alternative to ground-based methods, and has the potential to be incorporated into rangeland monitoring protocols.
pixmap: Bitmap Images (''Pixel Maps " ). R package version 0.4-11
  • R Bivand
  • F Leisch
  • M Maechler
Bivand, R., Leisch, F., Maechler, M., 2011. pixmap: Bitmap Images (''Pixel Maps " ). R package version 0.4-11. <http://CRAN.R-project.org/package=pixmap>.
Image analysis for automatic classification of Rumex obtusifolius in mixed grassland swards
  • S Gebhardt
  • W Kuehbauch
Gebhardt, S., Kuehbauch, W., 2006. Image analysis for automatic classification of Rumex obtusifolius in mixed grassland swards. J. Plant Dis. Prot. 113 (Sp. Iss. 20), 189-195.
Pixel Maps"). R package version 0.4-11
  • R Bivand
  • F Leisch
  • M Maechler
Bivand, R., Leisch, F., Maechler, M., 2011. pixmap: Bitmap Images (''Pixel Maps"). R package version 0.4-11. <http://CRAN.R-project.org/package=pixmap>.
Federal Biological Research Centre for Agriculture and Forestry
  • U Meier
Meier, U., 2001. BBCH monograph, second ed., 2001, Federal Biological Research Centre for Agriculture and Forestry.