Figure 3 - uploaded by Gerrit Polder
Content may be subject to copyright.
Example of a hyperspectral image data cube. Each pixels consists of a complete reflection spectrum at its position.  

Example of a hyperspectral image data cube. Each pixels consists of a complete reflection spectrum at its position.  

Source publication
Conference Paper
Full-text available
In the Interreg IV, EU project 'The healthy greenhouse' a new integral crop protection system is developed. Part of the project is the development of autonomous robots for monitoring individual plants. One of the sensors for monitoring is an application- specific multispectral camera for detection of fungal diseases. In this paper the development o...

Context in source publication

Context 1
... 2 shows a photograph of the hyperspectral imaging setup used in the laboratory. Figure 3 shows an example hyperspectral image cube. The spectral range is 400-1000 nm, in steps of 3.12 nm, resulting in 192 spectral bands. ...

Citations

... • The research in salient object detection in HSI has been mostly concerned with improving the results. There has Fig. 2 Hyperspectral image with spectral band frames and reflectance curves [12] been no attempt to include the processing time in the results. This paper is the first attempt to take the processing time into account and attempt to reduce it. ...
Article
Full-text available
The existing methods in salient object detection (SOD) in hyperspectral images (HSI) have used different priors like center prior, boundary prior to procure cues to find the salient object. These methods fail, if the salient object is slightly touching the boundary. So, we extrapolate boundary connectivity, a measure to check if the object touches the boundary. The salient object is obtained by using background and foreground cues, which are calculated using boundary connectivity and contrast map, respectively. Also, to reduce the information redundancy and hence time complexity, we select top three most informative bands using different feature selection and feature extraction algorithms. The proposed algorithm is tested on HS-SOD dataset. It is observed that the proposed algorithm performs better than the state-of-the-art techniques in almost all the metrics, such as Precision (0.57), Recall (0.46), f1f_1 score (0.51), CC (0.43), NSS (2.13), and MAE (0.09). In addition, we performed a comparative analysis of four different feature selection (MEV-SFS, OPBS) and feature extraction (PCA, MNF) algorithms in the context of SOD in HSI. We observed that feature selection algorithms are computationally efficient with OPBS and MEV-SFS taking about 7.98 and 8.34 s on average to reduce the feature space, respectively.
... An application-specific multispectral camera for the detection of fungi diseases is one of the cameras on the robot platforms. Cyclamen is used as a model crop, claims a study by Polder (2013). Cyclamen are especially susceptible to grey mould caused by Botrytis cinerea. ...
Chapter
Full-text available
The concept of agriculture has evolved over the past 20 years and now includes the processing, manufacturing, advertising, and supply of crops and livestock products. Currently, agricultural activities provide a basic means of subsistence, increase GDP, promote international trade, reduce unemployment, provide raw materials for other businesses, and generally improve the economy. Indian economy is significantly influenced by the agricultural sector. It employs more over 60% of the workforce and contributes roughly 17% of the nation's GDP. The modernization of agriculture has great potential to guarantee environmental safety, maximum productivity, and sustainability. The agricultural robot technology is an unavoidable requirement in agriculture industry, as its fundamental task is not only to solve the problem of less labor, precision, safety, comfort, and green operation, which is difficult to realize with traditional agricultural machinery and equipment, but also to fill the blank fields that many traditional types of agricultural machinery cannot fill although robots are still not as fast as humans in many cases. We'll probably witness even more fascinating advancements in this area in the years to come as AI continues to develop. Modern Agriculture: Exploring Current Trends 248 | P a g e
... Researchers have used hyperspectral sensors for ornamental crops, but mainly in laboratory applications due to their vulnerability in real-time field applications [43]. Polder et al. [48] identified Botrytis infected Cyclamen plants with selected features (bands) of 497, 635, 744, 839, 604, 728, 542, and 467 nm in a controlled greenhouse environment. Poona and Ismail [44] selected wavebands located across VIS, red edge, NIR, and SWIR regions to detect Fusarium circinatum infection in Pinus radiata seedlings at the asymptomatic stage. ...
Article
Full-text available
The ornamental crop industry is an important contributor to the economy in the United States. The industry has been facing challenges due to continuously increasing labor and agricultural input costs. Sensing and automation technologies have been introduced to reduce labor requirements and to ensure efficient management operations. This article reviews current sensing and automation technologies used for ornamental nursery crop production and highlights prospective technologies that can be applied for future applications. Applications of sensors, computer vision, artificial intelligence (AI), machine learning (ML), Internet-of-Things (IoT), and robotic technologies are reviewed. Some advanced technologies, including 3D cameras, enhanced deep learning models, edge computing, radio-frequency identification (RFID), and integrated robotics used for other cropping systems, are also discussed as potential prospects. This review concludes that advanced sensing, AI and robotic technologies are critically needed for the nursery crop industry. Adapting these current and future innovative technologies will benefit growers working towards sustainable ornamental nursery crop production.
... In spectroscopy, calibration techniques and surface reflectance correction, in general, use linear regression models, which is mainly a linear relation between the reflectance correction factor versus the wavelength variables. The calibration process is subject to two mathematical issues: handling several wavelengths and their correlation [32]. Figure 8 shows an example of hyperspectral image where the reflectance of the same spatial position (or a pixel) is shown along the different wavelengths. ...
... use linear regression models, which is mainly a linear relation between the reflectance correction factor versus the wavelength variables. The calibration process is subject to two mathematical issues: handling several wavelengths and their correlation [32]. Figure 8 shows an example of hyperspectral image where the reflectance of the same spatial position (or a pixel) is shown along the different wavelengths. ...
Article
Full-text available
The deployment of any UAV application in precision agriculture involves the development of several tasks, such as path planning and route optimization, images acquisition, handling emergencies, and mission validation, to cite a few. UAVs applications are also subject to common constraints, such as weather conditions, zonal restrictions, and so forth. The development of such applications requires the advanced software integration of different utilities, and this situation may frighten and dissuade undertaking projects in the field of precision agriculture. This paper proposes the development of a Web and MATLAB-based application that integrates several services in the same environment. The first group of services deals with UAV mission creation and management. It provides several pieces of flight conditions information, such as weather conditions, the KP index, air navigation maps, or aeronautical information services including notices to Airmen (NOTAM). The second group deals with route planning and converts selected field areas on the map to an UAV optimized route, handling sub-routes for long journeys. The third group deals with multispectral image processing and vegetation indexes calculation and visualizations. From a software development point of view, the app integrates several monolithic and independent programs around the MATLAB Runtime package with an automated and transparent data flow. Its main feature consists in designing a plethora of executable MATLAB programs, especially for the route planning and optimization of UAVs, images processing and vegetation indexes calculations, and running them remotely.
... Given the high cost and complexity that characterise hyperspectral cameras, research is oriented towards the building of cost-effective and accurate multispectral imaging systems, calibrated on specific wavebands, potentially dedicated to peculiar pathogens, and more suitable for the on-field applications [100,111]. For this purpose, starting from a laboratory-based hyperspectral system, the study in [112] selected eight bands suitable for the detection of grey mould in cyclamen plants, with the aim to develop a fast-multispectral camera. Multispectral sensors are comparable to hyperspectral ones, since both measure the up-coming reflected light from the leaf and/or canopy; however, a multispectral system often covers only three wavebands in the VIS (i.e., R, G, and B) in addition to NIR, allowing the calculation of some common vegetation indices, such as the NDVI. ...
Article
Full-text available
Ornamental plant production constitutes an important sector of the horticultural industry worldwide and fungal infections, that dramatically affect the aesthetic quality of plants, can cause serious economic and crop losses. The need to reduce the use of pesticides for controlling fungal outbreaks requires the development of new sustainable strategies for pathogen control. In particular, early and accurate large-scale detection of occurring symptoms is critical to face the ambitious challenge of an effective, energy-saving, and precise disease management. Here, the new trends in digital-based detection and available tools to treat fungal infections are presented in comparison with conventional practices. Recent advances in molecular biology tools, spectroscopic and imaging technologies and fungal risk models based on microclimate trends are examined. The revised spectroscopic and imaging technologies were tested through a case study on rose plants showing important fungal diseases (i.e., spot spectroscopy, hyperspectral, multispectral, and thermal imaging, fluorescence sensors). The final aim was the examination of conventional practices and current e-tools to gain the early detection of plant diseases, the identification of timing and spacing for their proper management, reduction in crop losses through environmentally friendly and sustainable production systems. Moreover, future perspectives for enhancing the integration of all these approaches are discussed.
... The spectral reflection data was collected using a hyperspectral line scan setup, similar to the one described in Polder and Young (2003) and Polder, Pekkeriet, and Snikkers (2013). The setup is shown in Fig. 1, and consisted of an ImSpector V10E spectrograph (SPECIM Spectral Imaging Ltd., Oulu, Finland) with a slit size of 30 mm, attached to a Photonfocus MV1_DV1320 camera (Photonfocus AG, Lachen, Switzerland) and a 25 mm lens. ...
Article
To handle surrounding objects, autonomous poultry house robots need to discriminate between various types of object present in the poultry house. A simple and robust method for image pixel classification based on spectral reflectance properties is presented. The four object categories most relevant for the autonomous robot PoultryBot are eggs, hens, housing elements and litter. Spectral reflectance distributions were measured between 400 and 1000 nm and based on these spectral responses the wavelength band with lowest overlap between all object categories was identified. This wavelength band was found around 467 nm with an overlap of 16% for hens vs. eggs, 12% for housing vs. litter, and less for other combinations. Subsequently, images were captured in a commercial poultry house, using a standard monochrome camera and a band pass filter centred around 470 nm. In 87 images, intensity thresholds were applied to classify each pixel into one of four categories. For eggs, the required 80% correctly classified pixels was almost reached with 79.9% of the pixels classified correctly. For hens and litter, 40–50% of the pixels were classified correctly, while housing elements had lower performance (15.6%). Although the imaging setup was designed to function without artificial light, its optical properties influenced image quality and the resulting classification performance. To reduce these undesired effects on the images, and to improve classification performance, artificial lighting and additional processing steps are proposed. The presented results indicate both the simplicity and elegance of applying this method and are a suitable starting point for implementing egg detection with the robot.
... The spectral reflection data was collected using a hyperspectral line scan setup, similar to the one described in Polder and Young (2003) and Polder, Pekkeriet et al. (2013). The setup is shown in Figure 4.4.1, and consisted of an ImSpector V10E spectrograph (Spectral Imaging Ltd.) with a slit size of 30 µm, attached to a Photonfocus MV1_DV1320 camera and a 25 mm lens. ...
... Four-band multispectral images with a NIR band added to the RGB bands are used for detection of the Tulip breaking virus (TBV) in tulip fields [4]. A customisable fast filter wheel multispectral camera was designed for detection of fungal diseases in the greenhouse [5]. ...
... The data on spectral reflection was collected using a hyperspectral line scan setup, based on the one mentioned in [16,17] and shown in Fig. 1. This setup used an ImSpector V10E spectrograph (Spectral Imaging Ltd.) with a slit size of 30 µm, attached to a Photonfocus MV1_DV1320 camera and a 25 mm lens. ...
Conference Paper
Full-text available
We present a simple and robust method for pixel segmentation based on spectral reflectance properties. Of four object categories that are relevant for PoultryBot, a mobile robot for poultry housings, the spectral reflectance was measured at wavelengths between 400 and 1000 nm. From this information, the distribution of reflectance values was determined for each combination of object category and wavelength band measured. From this, the wavelength band could be selected where the overlap between objects was lowest. This was found to be around 467 nm, with 16% overlap for chickens vs. eggs, 12% overlap for housing vs. litter, and lower overlap for other combinations. Images were taken with a standard monochrome camera and a band pass filter around 470 nm in a commercial poultry house, to test segmentation using this method. Preliminary results indicate that this method is a promising direction for future work.
Chapter
The paper raises the issue of lighting in industrial hyperspectral imaging. In the first part of the article, based on the literature review, a general overview of hyperspectral imaging is presented. Examples of applications of this unique technology are briefly listed. Despite significant progress in the area of data analysis and hyperspectral cameras, much remains to be done in terms of reliability and adaptation to industrial conditions. The authors emphasize the role of lighting in machine vision systems. Existing illumination solutions are described, along with their advantages and limitations. Due to specific requirements to be met in the case of hyperspectral imaging the authors propose two types of lighting devices. The prototype of a general usage halogen illuminator characterized by spatial uniformity and broadband spectral range is described in detail. The last part of the article presents the process of designing a dedicated illuminator along with the method of selecting relevant spectral bands for a particular application.