Crop acreage estimation is a key aspect to forecast crop production. Maize acreage estimation becomes more and more important because the fast production changes every year due to the dynamics of the prices. This paper focuses on maize acreage estimation in the North China Plain using ENVISAT MERIS and CBERS-02B CCD data of 2008. Firstly, adaptive maximum likelihood classification of CBERS-02B CCD images based on ground survey provided reliable maize area fraction image (AFI). CBERS derived AFIs (as reference AFI) were used to train a 3-layer back-propagation neural network, this was then used to the whole MERIS data to generate MERIS AFIs (AFIe). To estimate maize acreage, the maize AFI from MERIS was masked with cropland dataset and maize acreages were estimated by zonal statistic of maize AFI at district level. The statistical results were also modified using a non-arable coefficient to remove the effects of non-arable factors. The results showed a close relationship between estimated and statistical maize acreage (R2 ≈ 0.88). At province level, the estimation error is approximately 8%. This method is valuable for wide-scale, regional crop acreage estimation at the early stage of growing season. The study gives suggestions about high resolution image acquisition, spatial distribution and cropland datasets.Highlights► Valuable for wide-scale, regional crop acreage estimation at the early stage. ► Maize acreage estimation using ENVISAT MERIS and CBERS-02B CCD data. ► Crop acreage estimation combining high and coarse spatial resolution data.
Of late, a series of methods and tools have evolved in order to assist farmers in making decisions, based on the development of computers. In this paper, an attempt has been made to focus on simulation tools as a means to expand interactivity. The first part deals with the understanding of interactivity. Then the use of simulation as a support for interactivity is developed. In this endeavor, the evolution of understanding of the role played by simulation tools for farm management decision-making has been considered crucial, based on experiments carried out in an interactive manner with both farmers and consultants. Finally, we introduce and discuss numerous potential opportunities provided by the new techniques through automatic machine learning and multi-agent modeling.
A review is presented of several potentially useful applications of artificial neural networks (NN) to greenhouse climate control. Subjects covered are: Quasi-steady-state modelling, reduction (compression) of input and state vectors, NNs used as difference equations and replacing controllers (algorithms or humans) with NNs. In this context the strength of NNs is their flexibility to adapt to non-linear and non-physical data. Their main disadvantage is that their proper training requires large multi-dimensional sets of data to reduce the risk of extrapolation. Therefore, minimizing the dimensionality of the problem (both input and state vectors) becomes of paramount importance. Bottleneck NNs may be used for this purpose.
This paper describes a project to enhance farm level computer use education entitled The Year 2000 Computerized Farm. The farm management information system implemented at the Stiles Foundation Farm is described as well as its effect on farm profits. An evaluation of the short courses taught at the farm provide an insight as to the approach and results of that educational effort. The suggestion is made that other educational institutions could follow a similar approach for staff and user training.
A brief report is given of the field robot contest held in Wageningen, 5–6 June 2003. The experience of the student competitors sheds light on a number of issues that need to be addressed before autonomous vehicles will become a reality.
A preliminary study on the potential application of electronic tracking in poultry in vivo has been conducted. The experimental procedure for this study was based on previous in vitro findings (Fröschle et al., 2009) as part of the same research programme. The study consisted of two phases whereby an initial experiment using inkjet printing of 10 × 10 DataMatrix barcodes onto the beaks of broiler chickens in a live commercial setting has been carried out. Results demonstrated very poor percentage of readability after a short period of time. Barcodes deteriorated very rapidly and this was attributed to the physical effects on the barcodes of the actions of the chickens in a commercial environment, together with the inability of the ink to bond to the hard keratinous surface of the beak. In a subsequent part of the study, a number of commercially available ink types were screened, using a predetermined abrasion testing procedure, for their ability to bond to the beak and provide a readable barcode on the beaks following some predetermined graduated physical abrasion.
Contemporary Precision Livestock Technology in poultry production is very limited and does not meet European standards for traceability and Best Available Technology (BAT), as laid down in EN ISO 2205:2007 standards (2007) and the European Directive 2008/1/EC (2008). A worldwide occurrence of Avian Influenza additionally calls for a fraud-proof tagging device and source verification system for poultry and poultry products in order to complete partially existing documentary trails.During a preliminary laboratory trial, a procedure for the application of miniature linear and two-dimensional Data Matrix (DM) barcodes onto poultry beaks and legs through inkjet printing was set up and assessed. Results regarding the proportion of readability (p%), the standard error in readability (SE) and general statistics on the reading time were calculated. Tests for independence based on Chi-square and Pearson's were performed on the categorical data, to estimate the differences between proportions of readability of reading groups. The resulting data was used to define the optimal position of barcodes as well as the optimal reading mode of the barcode scanner to be used for further trials. As this experiment provided an estimate of readability of barcodes imprinted on chicken beaks and legs, it is intended to serve as a basis for sample size calculation for an ongoing live trial.
A tractor drawbar performance program that predicts the performance of two-wheel-drive (2WD) tractors for haulage as well as field operations for both bias-ply and radial-ply tyres is developed to meet user requirements in educational and research organizations. The program is written in Visual Basic programming language. The program provides an intuitive user interface by linking databases such as tractor specifications, tyre data, implement and trailer specifications and traction equation coefficients to predict the performance of a selected tractor model. The program has been proven to be user friendly and efficient for various field operations under frictional–cohesive soils.
Recent advances in sensor and wireless radio frequency (RF) technologies and their convergence with the Internet offer vast opportunities for development and application of sensor systems for agriculture. The objective was to create regional and on-farm sensor networks that provide remote, real-time monitoring and/or control of important farming operations that add value through improved efficiency and efficacy of targeted management practices. This paper describes hardware and software components of technologies we developed for regional and on-farm sensor networks and their implementation in two agricultural applications in Washington State, an agricultural weather network and an on-farm frost monitoring network. The regional sensor network consists of our AWN200 data logger equipped with a 900 MHz, frequency hopping, spread spectrum (FHSS) radio configured into master–repeater–slave network for broad geographic coverage. A single master is configured with multiple repeaters to provide a RF line-of-sight telemetry backbone network. Independent network backbones from disparate geographic regions are then aggregated in a central database via standard Internet protocols for further processing and dissemination. Software includes firmware to operate the data logger and radio telemetry aspects of the AWN200 in an agricultural weather network application called AgWeatherNet (http://www.weather.wsu.edu). The on-farm sensor network uses our SS100 radio/logger which includes a 900 MHz, FHSS radio, with software designed primarily for mobile, real-time farm operations and management applications. The network is deployed in a star topology in which a strategically placed base radio is responsible for network synchronization, data collection from remote stations within the network, and re-broadcasting collected data to roamer radio units attached to mobile computers and/or directly to the Internet. Client software, AgFrostNet, operating on a computer connected to a roamer, collects, manages, and display data in real-time. This software was designed specifically for air temperature monitoring during frost/freeze protection events. Both the regional AgWeatherNet WSN and the on-farm AgFrostNet networks were successfully implemented in Washington State. Problems encountered were mainly associated with power management under periods of low solar energy and with electrostatic discharge (ESD) damage to gallium-arsenide (GaAs) based transmit–receive switches in the radios during storms, a problem now corrected. Both systems have been made commercially available to growers via a novel arrangement between WSU and a local manufacturer.
In order to investigate and understand drift from field sprayers, a steady state computational fluid dynamics (CFD) model was developed. The model was developed in 3D in order to increase the understanding of the causes of drift: a deviation in the wind direction cannot be captured by a 2D approach, the wake behind a wind screen is not symmetrical, the effects of a changed nozzle orientation may not be symmetrical. The model's accuracy was validated with field experiments carried out according to the international standard ISO 22866. A field sprayer with a spray boom width of 27 m and 54 nozzles (Hardi ISO F110-03 at 3 bar) was driving at 2.22 m/s over a flat pasture. During the experiments the wind direction was perpendicular to the tractor track. The model explained the variation in drift replicates during each single field experiment through varying boom height (0.3–0.7 m), wind velocity (1.3–2.5 m/s), wind deviation (−18° to +18°) from the direction perpendicular to the tractor track and injection velocity of the droplets (17–27 m/s). Boom movements had the highest impact on the variations in drift values (deviations in drift deposits of 25%), followed by variation in wind velocity (deviations in drift deposits of 3%) and injection velocity of the droplets (deviations in drift deposits of 2.5%). Wind deviation from the direction perpendicular to the tractor track had a reducing effect on the drift values (deviations in drift deposits of 2%). Small variations in driving speed had little influence on drift values. Near drift (<5 m) is predicted well by the model but the increased complexity compromised the predictions at greater distances. The model will be further developed in order to improve far drift prediction. Dynamic simulations will be performed and the model for turbulent dispersion will be optimized. The model did not require calibration.
Several approaches have been reported previously to identify internal log defects automatically using computed tomography (CT) imagery. Most of these have been feasibility efforts and consequently have had several limitations: (1) reports of classification accuracy are largely subjective, not statistical; (2) there has been no attempt to achieve real-time operation; and (3) texture information has not been used for image segmentation, but has been limited to region labeling. Neural network classifiers based on local neighborhoods have the potential to greatly increase computational speed, can be implemented to incorporate textural features during segmentation, and can provide an objective assessment of classification performance. This paper describes a method in which a multilayer feed-forward network is used to perform pixel-by-pixel defect classification. After initial thresholding to separate wood from background and internal voids, the classifier labels each pixel of a CT slice using histogram-normalized values of pixels in a 3 × 3 × 3 window about the classified pixel. A post-processing step then removes some spurious pixel misclassifications. Our approach is able to identify bark, knots, decay, splits, and clear wood on CT images from several species of hardwoods. By using normalized pixel values as inputs to the classifier, the neural network is able to formulate and apply aggregate features, such as average and standard deviation, as well as texture-related features. With appropriate hardware, the method can operate in real time. This approach to machine vision also has implications for the analysis of 2D gray-scale images or 3D RGB images.
A 3D stand generator and visualization system was developed for generating a spatially explicit forest for central Appalachian hardwood forests. Spatial pattern of the stand generator was modeled and validated by characterizing a 75-year old central Appalachian mixed hardwood forest dominated by red oak, chestnut oak, red maple, and yellow poplar. All the trees larger than 12.7 cm diameter at breast height (DBH) were measured for DBH, total height, crown height and crown width along with their locations in thirty 0.162 ha plots. Stand attributes, i.e. species compositions, mean DBH, total height, basal area and volume in the generated stand never exceeded a difference of 10% from the actual stand attributes. The generated stand can thus be used as an alternate to time consuming manual measurement of spatial location of trees for related ecological studies and training purposes and to visualize the same in 3D perspectives. The results also indicated the potential of using this stand generator in simulating stand spatial patterns and generating stands in other regions with some modification in growth functions.
The most common method for in situ assessment of soil salinity, namely the electrical conductivity (EC) of the soil solution (ECw), is to measure the apparent electrical conductivity (ECa) and volumetric water content (θ) of the soil and apply measured or predicted ECa(ECw, θ) calibration curves. The water content and electrical conductivity of a soil solution are indeed the major factors affecting its apparent electrical conductivity, which justifies the assessment of salinity from apparent EC measurements. However, the ECa(ECw, θ) relationship depends on some additional soil and environmental attributes affecting the soil ECa. Non-spherical particle shapes and a broad particle-size distribution tend to decrease ECa, and when non-spherical particles have some preferential alignment in space, the soil becomes anisotropic, i.e., its ECa depends on the direction in which it is measured. The electrical conductance of adsorbed counterions constitutes a major contribution to the ECa of medium- and fine-textured soils, especially under conditions of low solution conductivity. In such soils and with such salinity levels, the temperature response of the soil ECa should be stronger than that of its free solution, and care should be taken when extrapolating from field-measured ECa values to obtain the ECa at a given temperature. The above-mentioned and other secondary findings should, on one hand, indicate some limitations for the application of existing ECa–ECw models, and, on the other hand, can serve as guidelines for further development of such essential models.
Dairy farmers using automatic milking are able to manage mastitis successfully with the help of mastitis attention lists. These attention lists are generated with mastitis detection models that make use of sensor data obtained throughout each quarter milking. The models tend to be limited to using the maximum or average value of the sensor data pattern, potentially excluding other valuable information. They often put cows on the lists unnecessarily, and their sensitivity for abnormal milk classification is too low for automated separation. Therefore, we analyzed sensor data patterns within quarter milkings in order to identify potentially predictive variables for abnormal milk and clinical mastitis classification. The data used in this study was obtained at a commercial dairy farm in Germany in September 2002, where a German Simmental herd was milked by a Lely Astronaut system. In total, 3232 quarter milkings from 63 cows were analysed; 94 quarter milkings were defined as milk with abnormal homogeneity and 270 as clinical mastitis. A data flow diagram was developed to systematically describe the steps involved in the transformation of within quarter milking measurements into variables that potentially predict abnormal milk and clinical mastitis. Three types of pattern descriptors were used: level, variability, and shape. In addition to using the absolute value of the pattern descriptor, the descriptors were considered relative to their expected value based on pattern descriptor values from previous milkings and from other quarters within the same cow milking. Using this method, potentially predictive variables were computed for electrical conductivity, the colours red, green and blue, a combination of colour sensors, and milk production. The importance of a variable in predicting abnormal milk and clinical mastitis was evaluated by computing correlation coefficients as well as information gain ratios. The most important variables came from the sensors for electrical conductivity, blue and green. Variables describing the variability and shape of the measurement patterns were as important as mean and maximum values, and should be included in future modelling. Also variables that are based on absolute values should be considered for future modelling. Results suggest that clinical mastitis and abnormal milk classification models may include similar predictive variables, but requirements for these models differ resulting in the need for different models. The schematic approach to developing potentially predictive variables will be helpful when exploring the usefulness of new sensors, researching other approaches to estimate expected values, and studying sensor data patterns in general.
Thinning is a silvicultural practice to improve tree growth and health. Thinning from below for the even-aged silviculture and thinning from above for the uneven-aged silviculture are the two mainly applied thinning practices. In forest management simulations, algorithms that describe which individual trees to be removed from a forest have developed in five growth simulators (Söderbergh and Ledermann, 2003). We have developed a shifting algorithm that determines the proportion of trees to be thinned from different diameter classes to complement the individual tree selection algorithms. Sampled (or mapped) tree diameters are grouped into diameter classes. Given the target thinning volume, the algorithm automatically computes the thinning rate in each of the diameter classes using the three-parameter Weibull distribution. The thinning rate is obtained by shifting the location parameter of an estimated Weibull distribution either to the right or to the left for thinnings from below and above, respectively. A modified bisection method is used to search for the new location parameter that yields the desired thinning volume. The proposed algorithm is demonstrated in examples by using experimental forest datasets. A stand-alone program called Weibull_thinning is downloadable at http://www.it.abo.fi/suswood/weibull_thinning/.Research highlights▶ A shifting algorithm to compute thinning rates by using the Weibull distribution for thinnings from below and above is presented. ▶ The shifting algorithm complemented the individual tree selection algorithms in published simulators. ▶ A modified bisection method is used to find the root, i.e., the target thinning volume.
The Russian wheat aphid (Diuraphis noxia (Mordvilko)) infests wheat (Triticum aestivum L.), barley (Hordeum vulgare L.), and other small grains and grasses. Russian wheat aphid infestations are unpredictable in time and space. In favorable conditions, Russian wheat aphid feeding can result in heavy damage to wheat and barley in a short period of time. A repetitive monitoring strategy that allows for rapid assessment of aphid infestation and damage over the growing season is critically needed. Tracking the irregular infestation patterns of Russian wheat aphid in order to optimize control efforts is central to the successful management of this aphid. One method that has been shown over a number of years to be useful for monitoring some insect outbreaks is to measure the light reflected by the infested canopy, plant, or leaf. Hence, this research was designed to investigate: (1) the potential use of remotely sensed data to discern and identify differences in spectral reflection patterns (spectral signatures) of winter wheat canopies with and without Russian wheat aphid infestation, and (2) the relationship between spectral indices and Russian wheat aphid abundance in wheat canopies growing in field conditions. Russian wheat aphid-infested wheat canopies had significantly lower reflectance in the near infrared region and higher in the visible range of the spectrum when compared with noninfested canopies. Linear regression analyses showed that there were varying relationships between Russian wheat aphid density and spectral vegetation indices, with coefficients of determination (r2) ranging from 0.91 to 0.01. These results indicate that remote sensing data have the potential to distinguish damage by Russian wheat aphid and quantify its abundance in wheat. However, success for Russian wheat aphid density estimation depends on the selection of spectral vegetation indices.
A particle acceleration device for implanting DNA material into plant cells using a timed burst of pressurized helium gas was developed. The device did not require the use of consumable materials (other than helium and microparticles), as used in some similar systems, and was thus easier and less expensive to operate. The system was designed to provide more precise control of the duration of the helium burst and of pressure differentials to ensure more reproducible results. Safety interlock circuitry was incorporated. The device was found to be efficient for delivery of DNA into alfalfa and soybean cell cultures.
Automatic monitoring of animal behavior in livestock production opens up possibilities for on-line monitoring of, among others, oestrus, health disorders, and animal welfare in general. The aim of this study is to use time series of acceleration measurements in order to automatically classify activity types performed by group-housed sows. Extracts of series collected for 11 sows are associated with 5 activity types: feeding (FE), rooting (RO), walking (WA), lying sternally (LS) and lying laterally (LL). A total of 24 h of three-dimensional series is used. One univariate model and four multivariate models are used to describe all five activity types. Three multivariate models differ in their variance/covariance structure; a fourth alternative multivariate model (MU) simply combines the 3-axes of the univariate model, assuming independence. For each model, the activity-specific parameters are estimated using the EM algorithm. The classification method, based on a Multi-Process Kalman Filter provides posterior probabilities for each of the 5 activities, for a given series. For the univariate model, LL is the activity which is best recognized by the 3-axes; FE, RO and WA are best recognized by one particular axis; LS is poorest recognized. The average results are improved by using all four types of multivariate models. The percentages of activity recognition are similar among the multivariate models. By grouping the activity types into active (FE, RO, WA) vs. passive (LS, LL) categories, the method allows to correctly classify 96% of the active category and 94% of the passive category.
Cattle behavior is potentially a valuable indicator of health and well-being; however, natural movement patterns can be influenced by the presence of a human observer. A remote system could augment the ability of researchers, and eventually cattle producers, to monitor changes in cattle behavior. Constant video surveillance allows non-invasive behavior monitoring, but logging the movement patterns on individual animals over long periods of time is often cost prohibitive and labor intensive. Accelerometers record three-dimensional movement and could potentially be used to remotely monitor cattle behavior. These devices collect data based on pre-defined recording intervals, called epochs. Our objectives were to (1) determine if accelerometers can accurately document cattle behavior and (2) identify differences in classification accuracy among accelerometer epoch settings. Video-recorded observations and accelerometer data were collected from 15 crossbred beef calves and used to generate classification trees that predict behavior based on accelerometer data. Postural orientations were classified as lying or standing, while dynamic activities were classified as walking or a transition between activities. Video analysis was treated as the gold standard and logistic regression models were used to determine classification accuracy related to each activity and epoch setting. Classification of lying and standing activities by accelerometer illustrated excellent agreement with video (99.2% and 98.0% respectively); while walking classification accuracy was significantly (P < 0.01) lower (67.8%). Classification agreement was higher in the 3 s (98.1%) and 5 s (97.7%) epochs compared to the 10 s (85.4%) epoch. Overall, we found the accelerometers provided an accurate, remote measure of cattle behavior over the trial period, but that classification accuracy was affected by the specific behavior monitored and the reporting interval (epoch).
Theoretically, using information about crop growth would allow the extension of present greenhouse control strategies towards a truly economic optimal control strategy. A brief survey is given of developments in the scientific literature. A full solution would require to consider the long-term crop development as well as all relevant short-term dynamics of the crop, the greenhouse and the external weather. Obstacles for the acceptance of such solutions are briefly discussed. One of the key factors is the lack of reliable crop development models for the wide variety of crops encountered in practice, and the need to leave part of the decision freedom in the hands of the grower. An analysis is given of simplified approaches resulting from integrating the crop equations over a day or more. The temperature integral concept, a specific example of such approach, is gaining popularity, despite the fact that it lacks exploitation of knowledge about the fast crop responses. The discussion leads to the concept of separation of responsibilities, where the short-term effects, including photosynthesis and evapo-transpiration, are handled by an automated model-predictive optimal controller, while the long-term effects are left to the grower, with support from a flexible decision support system based on crop models whenever they become available.
A relational database was developed for the agricultural chemical use data collected by the US Department of Agriculture, National Agricultural Statistics Service since 1990. coldfusion Markup Language was used for the client-side interface and server side process programming. The database is accessible from the Web at URL: http://www.pestmanagement.info/nass. Users can obtain information about agricultural chemical use in the database by search of crop, year, region, and active ingredient. Various agricultural chemical usage statistics are provided as Web tables, dynamically generated US maps, charts and graphs, and downloadable Excel files. We used a centralized software architecture in this project, which is suitable for projects with moderate programming complexity. A distributed approach might be more appropriate for the more complex projects. The current database information, spanning 1990–2001, will be augmented in the future, possibly using an automated updating scheme.
Writing down mathematical models of agricultural greenhouses and regulating
them via advanced controllers are challenging tasks since strong perturbations,
like meteorological variations, have to be taken into account. This is why we
are developing here a new model-free control approach and the corresponding
intelligent controllers, where the need of a good model disappears. This
setting, which has been introduced quite recently and is easy to implement, is
already successful in many engineering domains. Tests on a concrete greenhouse
and comparisons with Boolean controllers are reported. They not only
demonstrate an excellent climate control, where the reference may be modified
in a straightforward way, but also an efficient fault accommodation with
respect to the actuators.
This study compared two alternative techniques for predicting forest cover types from cartographic variables. The study evaluated four wilderness areas in the Roosevelt National Forest, located in the Front Range of northern Colorado. Cover type data came from US Forest Service inventory information, while the cartographic variables used to predict cover type consisted of elevation, aspect, and other information derived from standard digital spatial data processed in a geographic information system (GIS). The results of the comparison indicated that a feedforward artificial neural network model more accurately predicted forest cover type than did a traditional statistical model based on Gaussian discriminant analysis.
A low-cost, non-differentially corrected hand-held GPS receiver was tested on an industrial peat production bog. A correction procedure (‘pseudo-differential correction’) was derived that corrected data points to the nearest position on a line defining the centre of each 15-m wide field. The result was a corrected log of track points for each field for all points lying along the field. It was found that the mean orthogonal distance from a field centreline was linearly correlated with mean uncorrected GPS data error (r2=0.99) such that as GPS error increased so the accuracy obtained by correction decreased. For a signal with a mean uncorrected error of ∼30 m it was possible to reduce the error to ∼12 m. The results are discussed within the design requirements of a precision peat production system for peat energy. It is concluded that low-cost GPS could be used without differential correction as part of a precision peat production system because over 80% of the time positional error could be constrained to within 15 m. When compared with the perceived patterns of variability and the 30-m resolution of Landsat imagery which can be used for making application maps, this is acceptable.
Soil apparent electrical conductivity (ECa) has been used as a surrogate measure for such soil properties as salinity, moisture content, topsoil depth (TD), and clay content. Measurements of ECa can be accomplished with commercially available sensors and can be used to efficiently and inexpensively develop the dense datasets desirable for describing within-field spatial variability in precision agriculture. The objective of this research was to investigate accuracy issues in the collection of soil ECa data. A mobile data acquisition system for ECa was developed using the Geonics EM381 sensor. The sensor was mounted on a wooden cart pulled behind an all-terrain vehicle, which also carried a GPS receiver and data collection computer. Tests showed that drift of the EM38 could be a significant fraction of within-field ECa variation. Use of a calibration transect to document and adjust for this drift was recommended. A procedure was described and tested to evaluate positional offset of the mobile EM38 data. Positional offset was due to both the distance from the sensor to the GPS antenna and the data acquisition system time lags. Sensitivity of ECa to variations in sensor operating speed and height was relatively minor. Procedures were developed to estimate TD on claypan soils from ECa measurements. Linear equations of an inverse or power function transformation of ECa provided the best estimates of TD. Collection of individual calibration datasets within each surveyed field was necessary for best results. Multiple measurements of ECa on a field were similar if they were obtained at the same time of the year. Whole-field maps of ECa-determined TD from multiple surveys were similar but not identical. There was a significant effect of soil moisture and temperature differences across measurement dates. Classification of measurement dates as hot vs. cold and wet vs. dry provided TD estimations nearly as accurate as when individual point soil moisture and temperature data were included in the calibration equation.