Hyperspectral imaging in crop fields: precision agriculture

Chapter 3.3

Hyperspectral imaging in crop fields: precision agriculture Daniel Caballero,a, b Rosalba Calvinic and Jose´ Manuel Amigoa, * a

Professor, Ikerbasque, Basque Foundation for Science; Department of Analytical Chemistry, University of the Basque Country, Spain; Chemometrics and Analytical Technologies, Department of Food Science, University of Copenhagen, Denmark; bComputer Science Department, Research Institute of Meat and Meat Product (IproCar), University of Extremadura, Ca´ceres, Spain; c Department of Life Sciences, University of Modena and Reggio Emilia, Modena, Italy *Corresponding author.

1. Crop fields, precision agriculture, and chemometrics. A general overview As it is well known, hyperspectral imaging (HSI) and multispectral imaging (MSI) were originally developed for remote sensing and earth observation via satellite. With the new technological advances in sensing and data analysis, they have gained wide recognition as nondestructive and fast analytical tools for quality assessment of crop fields. Technological advances are deeply modifying agriculture and future scenarios including the possibility for farmers and agronomists to remotely manage their field, thanks to real-time images providing relevant information about crop disease, nutrient status, pest infections [1], detection of contaminants [2e4], harvest identification [5e7], constituent analysis [8,9], or quality evaluation [10]. This originated the so-called precision agriculture. A modern way of growing crops that allows the farmer to keep direct real-time control of the quality parameters of their fields. Precision agriculture implies the use of advanced cameras and sensors that normally are implemented in satellites, different types of aerial vehicles, and, nowadays, even terrestrial vehicles. The cameras, normally denoted as HSI or MSI need, of course, to be integrated with the most efficient algorithms in order to give a reliable response to the farmer at real time. The distinction between HSI and MSI is usually defined as function of number of spectral bands measured [11]. Additionally, they can also be defined as a function of the wavelength region used to acquire the Hyperspectral Imaging. https://doi.org/10.1016/B978-0-444-63977-6.00018-3 Copyright © 2020 Elsevier B.V. All rights reserved.

453

454 SECTION j III Application fields

images. Normally, visible (from 350 until 800 nm, approximately) and nearinfrared (NIR) (from 800 until 2600 nm, approximately) are the most suitable spectral regions. Remote sensing is continuously expanding due to the advances in sensors for imaging spectroscopy, allowing the improvement of portability of radiometers and spectrophotometers in satellites, planes, or other aerial vehicles (e.g., drones or helicopters) [12,13]. In the last 25 years, optical sensor technology has suffered a rapid development to obtain more high temporal and spatial resolution remote sensing images. A demonstration of this is the high amount of satellites existing nowadays that are able to collect different spectral and spatial information (e.g., Sentinel-2, Landsat-8, RapidEye, SPOT-6, GeoEye-1, Huanjing-1) [14]. One of them, Hyperion, has a sensing spectrum of 400e2500 nm with a 10-nm resolution in 220 bands. It covers a swath width of 7.5 km at a spatial resolution of 30 m [15]. In some other examples, Kross et al. [16] estimated biomass of corn using RapidEye NIR-HSI data. Li et al. [17] comparatively analyzed Gaofen-1, Huanjing-1, and Landsat-8 NIR-HSI for estimating the area of winter wheat. Clevers et al. [18] estimated the quality parameters of a potato crop with different fertilization levels using Sentinel-2 satellites NIR-HSI. This study shows that the difference vegetation index calculated using bands at 10 m spatial resolution can be used for estimating the crop quality parameters [18]. The main advantage of using aerial vehicles is the quick and repeated acquisition capability. These systems acquire data flying at low heights and provide very high resolution imagery at, e.g., farm scale. Nowadays, it is possible to capture highly accurate images of individual fields covering up to hundreds of hectares in a single flight by drone without the cost and hassle of manned services at a far greater resolution than satellite imagery offers [19]. Some examples are the estimation of mycotoxin-producing pathogens in some crop fields, such as maize, wheat, sugarcane, and sugar beet [20,21]. Fig. 1 shows the quantitative result of an NIR-HSI discriminating between the different crop fields. The green areas show high density of potatoes, the yellow zones show medium density, and brown zones are areas without potatoes. This great capability of collecting real-time spectral data is not exempt of drawbacks. One of them is the physical complexity of the data acquired. Fast computers, sensitive detectors, and large data storage capacities are needed for analyzing hyperspectral data [24]. In addition, one of the challenges that researchers have to face is finding ways to program hyperspectral satellites to sort through data on their own and transmit only the significant bands of the images. For that, the last great advances in remote sensing and computer technology are revolutionizing the way that data are collected and analyzed [25e28], in particular, the incorporation of latest-generation sensors to different platforms for earth observation [29]. Thus, the development of computationally efficient techniques for transforming the massive amount of data into scientific understanding is critical [30,31].

Hyperspectral imaging in crop fields: precision agriculture Chapter j 3.3

455

FIGURE 1 Precision agriculture among different crop fields. From P.P.J. Roosjen, B. Brede, J.M. Suomalainen, H.M. Barthoolomeus, L. Kooistra, J.G.P.W. Clevers, Improved estimation of leaf area index and leaf chlorophyll content of a potato crop using multi-angle spectral data e potential of unmanned aerial vehicle imagery, International Journal of Applied Earth Observation and Geoinformation, 66 (2018) 14e26.

The utilization of high-performance computing systems in remote sensing applications has become more and more widespread [32]. The development of new multiprocessor systems for extracting image information allows utilizing heterogeneous data computing resources from different sources [33e36]. Thus, although remote sensing data processing algorithms generally map quite nicely to multiprocessor systems made up of clusters of networks of CPUs, these systems are expensive and difficult to adapt on board for remote sensing data processing. In this regard, the emergence of specialized hardware devices as field programmable gate arrays [37] or graphics processing units [38] exhibit the potential to bridge the gap between the real time analysis and remote sensing data acquisition. These aspects are of great importance in the definition of remote sensing missions, in which the payload is an important parameter. The spectral signatures collected are also affected by several issues. One of them is the fact that pixels contain mixed spectral information (one pixel might contain spectral information of many plants, species, or even more than one vegetation type [39,40]). Moreover, depending of the configuration, atmospheric and illumination effects hamper the straightforward acquisition of

456 SECTION j III Application fields

straight analytical responses. Even though the detection of these issues should be done automatically in an operational environment resulting in a reliable information [41], there is a strong synergy between remote sensing and chemometrics [42]. Indeed, the success of the remote sensing cannot be understood without referring to the implementation of powerful algorithms to handle all data generated for each image. Nansen et al. [42a] have evaluated the effect of different lighting conditions in the identification of sugarcane borer (Diatraea saccharalis) infestation in individual maize plants. The experiments were conducted by imaging the same plants with a vis-NIR hyperspectral camera under two different lighting conditions: in a shaded greenhouse with controlled illumination and under direct sunlight. Notwithstanding the use of the same internal standards for image correction, classification models developed under one lighting condition were not successfully transferred the other lighting condition. These results suggest the need of advanced correction methods able to minimize the changes due to variations in the illumination conditions, leading to the calculation of more robust models. Chemometrics is a well-known discipline that allows extracting information initially hidden from large data sets in a multivariate way. Many reviews point out the main multivariate or statistical methods to be applied for different purposes [11,43e45]. The interest in chemometrics techniques has increased because of the rapidly decreasing cost of large storage devices and growing ease in data collection over networks. Other factors include, the development of robust and efficient algorithms to process these data, and the increase in computing power, enabling the use of intensive computational methods for data analysis [46]. In the last years, new algorithms have performed relatively well in the preprocessing and processing remote sensing images. What concerns to preprocessing, normal operational atmospheric correction procedures involve, for instance, the correction of the effects of aerosols, clouds, sun glint, and adjacency effects [47] even in a simple correction such as the dark object corrections [48,49]. Once those issues are minimized, deep learning [50,51], multiobjective optimization [52e54], and the evolutionary multitasking [55e57] have largely demonstrated their utility. Fig. 2 is an example of a typical multitarget classification in which different methods are used to discriminate between river, crop fields, vegetation, and buildings. This chapter provides a review of the scientific literature related with the application of remote sensing on crop fields and their applications on vegetation and ecosystems. The main advantages and constraints are also shown and studied in order to define future challenges.

2. Detection of contaminants Crop fields are in the focus of plant research worldwide given their important role in human food production and animal husbandry [58e62]. Some

Hyperspectral imaging in crop fields: precision agriculture Chapter j 3.3

457

FIGURE 2 Clustering results by new methods. (A) FCM, fuzzy c-means; (B) FCIDE, automatic fuzzy clustering using an improved differential evolution algorithm; (C) AMASFC, adaptive memetic fuzzy clustering algorithm with spatial information; (D) AFCMDE, automatic fuzzy clustering method based on adaptive multi-objective differential evolution; (E) AFCMOMA, adaptive multi-objective memetic fuzzy clustering algorithm. From Y. Zhong, A. Ma, Y.S. Ong, Z. Zhu, L. Zhang, Computational intelligence in optical remote sensing image processing, Applied Soft Computing 64 (2018) 75e93.

applications of HSI and MSI revised in this chapter are the detection of contaminants and heavy metals in cereal grains, the use of water assessments in the crop fields, the health applications on the crop fields, and the applications on preharvest products. The spectral behavior of heavy metals in soils or plants is commonly measured using NIR-HSI [63,64]. The main working procedure is to create a comprehensive and reliable database of many different soil or vegetation samples and then use those measurements as references for a library to compare the pixels measured by the hyperspectral or multispectral camera [65e67]. Since multiple metals are likely to be copresent in the soil, the exact spectral response of a given metal is usually isolated through simulations or using the well-known spectral unmixing methodologies [68]. Table 1 presents the feasibility of detection of heavy metals from remotely sensed imagery in practice samples. High correlation coefficients (R > 0.970) were reached for

458 SECTION j III Application fields

TABLE 1 Detection of heavy metals at crop fields by using remote sensing. Imagery

Detected metals

Targets

Authors

Satellite

Ni, Cu, and Cr

Soil fields

[69]

Satellite

Zn

Soil fields

[70]

Satellite

Cu and Cd

Rice

[71]

Satellite

Cu and Ni

Leaves

[72]

Plane

Pb, Zn, Cu, Cd, and Mn

Soil fields

[73a] [73] [74] [75]

Plane

As

Leaves

[76]

Plane

Cu and Zn

Soil fields

[77]

Plane

Cd and Zn

Soil fields

[78]

the metals presented in the table [74]. This high correlation between heavy metal content determined in a laboratory and spectral reflectance suggests that heavy metal concentrations can be retrieved from spectral reflectance (NIR-HSI) at high accuracy. Rathod et al. [76] quantified As in barley leaves with a significant correlation with chlorophyll and water stress indices. Residual heavy metal contamination has been estimated in an area affected by mining [73] by using NIR-HSI. Choe et al. [73a] have mapped heavy metal distribution in stream sediments and salt efflorescence on mine waste by NIR-HSI. Riaza et al. [75] found that subtle mineralogical changes can be discerned from the spectral responses (NIR-HSI). In other studies, herbaceous, shrub biomass, Cd, and Zn contents were determined from species-level vegetation map (NIR-HSI) [77]. Highresolution sensors can resolve the subtle disparity between the spectral responses (NIR-HSI) of intact and affected plants for estimating Cd, Cu, and Zn content [78]. Hyperion has been used to map the spatial distribution (NIR-HSI) of Zn based on difference moisture stress index [70]. The spatial distribution (NIR-HSI) of Cu and Cd in rice was mapped at three concentration levels at accuracy ranging from R2 ¼ 0.69 to 0.72, demonstrating the feasibility of large-scale monitoring rice contamination by heavy metals [71]. In addition, Zhou et al. [72] detected the levels of Cu and Ni in soil fields by using multilevel target updates method. Fig. 3 shows an example with crop fields containing different concentrations of heavy metals, being highlighted in blue and light green the zones with

Hyperspectral imaging in crop fields: precision agriculture Chapter j 3.3

459

FIGURE 3 Heavy metal pollution in rice in all of the research area. From X. Li, J. Li, X. Liu, Collaborative inversion heavy metal stress in rice by using two-dimensional spectral feature space based on HJ-1 A HIS and radarsat-2 SAR remote sensing data, International Journal of Applied Earth Observation and Geoinformation 78 (2019) 39e52.

a high concentration of heavy metals, while the dark green areas show low concentration of heavy metals.

3. Water stress and health status The level of technology used in hyperspectral sensors has been significantly increasing during the last years, specially to analyze the water on crop fields [80]. This is an increasing trend, mainly due to advances in ground remote sensing. The healthy plants absorb most of the visible lights that receive them while reflecting a large portion of NIR light. The dead plants absorb the visible and NIR lights, while the stressed vegetation absorb most of red and blue channels of light, reflecting the green channel of light and a large portion of NIR light (Fig. 4). These are good indicators of nitrogen content, biomass, chlorophyll, leaf area index at the leaf and canopy level for different species and have a good correlation to soil water content. A very simple but efficient way of using the spectral information for having a quick vision of the health status in vegetation is the well-known vegetation indices. Among them are the Normalized Difference Vegetation Index (NDVI)

460 SECTION j III Application fields

FIGURE 4 Light reflections in healthy, stressed, and dead leaves. The graphs below indicate the relative reflection in the blue (B), green (G), red (R), and near-infrared (NIR) channels.

(Eq. 1), Green Normalized Difference Vegetation (green NDVI-I, Eq. (2), and green NDVI-II), Water Band Index (WBI) (Eq. 3), Soil Adjusted Vegetation (SAVI), Photochemical Reflectance (PRI), Red-edge Vegetation Stress (RVSI), Modified Chlorophyll Absorption in Reflectance (MCARI), and Visible Atmospherically Resistant Index (VARI). Some examples are shown in the following equations: NDVI ¼

ðR800  R640 Þ ðR800 þ R640 Þ

green NDVI  I ¼ WBI ¼

ðR550  R640 Þ ðR550 þ R640 Þ

R950 R900

(1)

(2)

(3)

where R denotes the value of the radiance at the corresponding wavelength (in nanometers). Table 2 shows some successful case studies related to reflectance indices of water stress assessment in different irrigation treatments applied on open crop field for different vegetal species. One of the main advantages of the vegetation indexes is that they are sensitive to water stress conditions through depoxidation state of xanthophylls. It is important to control the plants under water stress, since when the demand

Hyperspectral imaging in crop fields: precision agriculture Chapter j 3.3

461

TABLE 2 Reflectance indices for water stress assessment from different species by using hyperspectral imaging and remote sensing. Imagery

Reflectance indices

Species

Authors

Satellite

Single band

Green beans

[81]

Satellite

Difference vegetation index

Corn, spinach

[82]

Satellite

Simple ratio

Apple, potatoes

[83]

Satellite

Simple ratio

Cotton

[84]

Satellite

Moisture stress index

Vineyard

[85]

Satellite

Moisture stress index

Wheat

[86]

Satellite

Normalized difference vegetation index

Barley

[87]

for water exceed the water supply in the root zone, the plants remove the capability to transport the water from the root zone to the atmosphere, and the photosynthesis rate is decreased. Some studies [88,89] showed that the reflectance in the green and red bands under water stress is increased due to leaf chlorophyll concentration reduction (Fig. 5). In Fig. 5, the variability of water concentration in one crop field can be seen. Red pixels represent plants with a high water content and green, plants that suffer low water. This variability could be related with the ability of the soil in the different sections of the crop fields. The main disadvantage is that the intensity of the changes varies as a function of the environmental conditions, and the correlation with water content is indirect [81,83,84]. For the moisture stress index, the main advantage is the high correlation with the moisture content and plant water content,

FIGURE 5 Levels of water concentration in crop fields. (A) true colour image of a NVT (National Variety Trials) wheat field after sowing (the red rectangles are the experimental field boundaries); (B) image resulting from dividing first and second principal component images. From Y. Sadeh, X. Zhu, K. Chenu, D. Dunkerley, Sowing date detection at the field scale using CubeSats remote sensing, Computers and Electronics in Agriculture 157 (2019) 568e580.

462 SECTION j III Application fields

and the main disadvantage is the variability among replicates with leaves at the same level of water stress [85,86].

4. Crop diseases The International Panel on Climate Change (IPCC) has reported that the most hazardous manifestation of climate change is through increased temperature, heat waves, and prolonged droughts, which in turn cause sever fluctuations in crop fields productions [91]. One of the main reasons for this is related to the infection of plants by insects and pests, which strongly affects crop yields and causes huge economic losses and the impact in the shifting in the occurrence of diseases. Despite these losses, the greatest challenge is the lack of knowledge regarding the probability of occurrence, distribution, and diseases spread, as well as the severity of their effect on crop fields [92]. Over the past decades, MSI and HSI cameras in the visible and NIR regions mounted over airplanes and drones proved to be an effective tool for a fast and reliable assessment of crop infection. In this manner, the time needed to detect the presence of possible crop infections is strongly reduced and, at the same time, it is also possible to adopt targeted pest management strategies by optimizing the amount of pesticides to be used for the specific need of the field [93]. Apart from the lightning problems mentioned before, a key issue of remote sensing in pest management is related to the low spatial resolution of hyperspectral cameras compared to the dimensions of insects. Indeed, insect crop infestations are not identified by detecting the insects themselves, but by identifying biological and physiological changes induced in plant leaves. Therefore, spatial resolution of remote sensing data should be high enough to allow the detection of pest-induced damage at least at the plant level and ideally at the leaf level [93]. For these reasons, aerial images usually provide more accurate results than satellite images [94]. Several studies have demonstrated that insect herbivory adversely affect the ability of plants to perform photosynthesis [95e97]. Consequently, the reflectance profile of leaves will change accordingly due to the lower light absorption by leaf pigments, such as chlorophylls and carotenoids [98]. Elliot et al. [99] used airborne MSI to study the damages induced by Russian wheat aphids (Diuraphis noxia) to wheat plants; in particular the authors were able to relate vegetation indexes calculated from multispectral data with the amount of infected plants. Reisig and Godfrey [100] demonstrated the possibility of using NIR aerial images to discriminate cotton plants infested by aphids (Aphis gossypii Glover) and spider mite (Tetranychus spp.) from uninfected plants. Moreover, insect infestations can cause major damages in forests, and also in this field of application remote sensing resulted to be an effective tool for pest management [101]. As mentioned above, the underlying principle of hyperspectral remote sensing for pest identification is to recognize the reflectance changes of

Hyperspectral imaging in crop fields: precision agriculture Chapter j 3.3

463

leaves caused by insect damages. However, leaf damage can also be caused by other biotic or abiotic stressors, and therefore it is important to correctly identify the reason of the damage. To this aim, Backoulou et al. [102] demonstrated that spatial information contained in multispectral images can be used to identify spatial patterns specific to leaf damages caused by insects. The same authors applied this strategy to quantify stress in wheat fields in order to distinguish between damages caused by D. noxia from those caused by other stress factors such as drought and adverse agronomic conditions [103,104]. Some methods have been developed for assessing the spatial spread of diseases in crop fields. Climate models have been used to map the potential distribution of pathogens on basis of the fundamental niche concept [105]. This approach provides a suitable tool for exploring the potential of diseases spreading into new areas. Diseases for plants increase reflectance in the visible range, particularly in the red absorption feature (670 nm) [106]. Fluorescence (450e550 nm) and heat sensing (8000e14,000 nm) present a potential avenue for quantifying the photochemical efficiency of plants, an indicator of their health status [92]. Fig. 6 shows multispectral images from maize acquired with different wavelengths, showing the vitality of the leaves, the blue color represents healthy leaves and the purple color represents lesioned leaves. The increasing availability of small, high-resolution spatial and spectral sensors has improved the operational capabilities of remote sensing through unmanned aerial vehicles (drones) mounting spectral sensors for crop fields monitoring at the farm scale [108]. The measuring range within the on-the-go proximal sensing techniques varies from the visible to NIR [13]. The technique has been successfully applied which has the ability to provide several soil characteristics [109].

FIGURE 6 Interpretation of hyperspectral images of maize for detecting healthy and dead leaves of maize. From J. Behmann, J. Steinru¨cken, L. Plu¨mer, Detection of early plant stress responses in hyperspectral images, ISPRS Journal of Photogrammetry and Remote Sensing 93 (2014) 98e111.

464 SECTION j III Application fields

5. Pre-harvest applications Consumers would pay more for prime food and agricultural products with high quality and safety guaranteed than for products with a low quality and from unknown origin. In this sense, HSI has been used for sample classification and quality grading; detection; and visualization of chemical attributes of fruits, vegetables, and other products on the preharvest stage. Table 3 summarizes the main applications of HSI in the evaluation of agricultural products. As shown in Table 3, the quality parameters, which have been investigated for fruits and vegetables, mainly include defects [5,110,120], firmness [111], color [119], pit detection [112], bruise [114], and additional quality parameters of fruits and vegetables [113,115,116,118]. Toxigenic fungi [117] have grown in grain, and they are toxics for animals and humans. In addition, there are many works which have been conducted for the grain analysis, by using NIR-HSI [121], and for predicting the quality of potatoes [122] or kiwis [123].

6. Conclusions and further challenges The advantages of the use of the remote sensing and the correct image analysis techniques have provided an opportunity to develop a vast amount of applications in crop fields. Further research should focalize on the use of new image

TABLE 3 Main applications of hyperspectral imaging in quality evaluation of agricultural and preharvest products. Imagery

Product

Application

Authors

Satellite

Apple

Defects

[110]

Satellite

Blueberry

Firmness

[111]

Satellite

Cherry

Pit detection

[112]

Satellite

Grape

Quality evaluation

[113]

Satellite

Orange

Defects

[5]

Satellite

Peach

Bruise

[114]

Satellite

Strawberry

Quality evaluation

[115]

Satellite

Corn

Quality evaluation

[116]

Satellite

Wheat

Fungus detections

[117]

Satellite

Cucumber

Quality evaluation

[118]

Satellite

Mushroom

Color

[119]

Satellite

Tomato

Defects

[120]

Hyperspectral imaging in crop fields: precision agriculture Chapter j 3.3

465

analysis techniques to obtain the maximum potential for mapping and monitoring the crop fields, for health applications or to detect contaminations on the harvests. Thus, further research also should investigate the use of alternative sensing techniques and its potential in validating findings derived from the use of more affordable HSI acquisition techniques. With the increasing use of optical remote sensing images with a spatialtemporal resolution, computational intelligence techniques have been widely used in remote sensing image processing. In some applications, the performance of these methods is affected by the errors made of the remote sensing data. Therefore, the use of proper multivariate methods and autolearning methods could obtain better performance in these types of tasks. Many computational intelligence techniques have been introduced to remote sensing image processing. However, it is worth noting that the massive amounts of images with a high spatial-spectral-temporal resolution, the diverse descriptors that are generated, and the ratio with the images are acquired and updated could result in an imperative need for rapid advances in technologies based on computational intelligence.

References [1] [2]

[3]

[4]

[5] [6]

[7]

[8]

[9]

P.L. Hatfield, P.J. Pinter Jr., Remote sensing for crop protection, Crop Protection 12 (6) (1993) 403e413. C.D. Everard, M.S. Kim, H. Lee, A comparison of hyperspectral reflectance and fluorescence imaging techniques for detection of contaminants on spinach leaves, Journal of Food Engineering 143 (2014) 139e145. B. Kuswandi, D. Futra, L.Y. Heng, Nanosensors for the detection of food contaminants, in: A.E. Oprea, A.M. Grumezescu (Eds.), Nanotechnology Applications in Food: Flavor, Stability, Nutrition and Safety, Academic Press, New York, New York, U.S.A, 2017, pp. 307e333. K.T. Sanders, S.F. Masri, The energy-water agriculture nexus: the past, present and future of holistic resource management via remote sensing technologies, Journal of Cleaner Production 117 (2016) 73e88. J. Li, X. Rao, Y. Ying, Detection of common defects on oranges using hyperspectral reflectance imaging, Computers and Electronics in Agriculture 78 (2011) 38e48. D. Wu, D.W. Sun, Advanced applications of hyperspectral imaging technology for food quality and safety analysis and assessment: a review e part II: applications, Innovative Food Science and Emerging Technologies 19 (2013) 15e28. K. Yu, Y. Zhao, X. Li, Y. Shao, F. Zhu, Y. He, Identification of crack features in fresh jujube using Vis/NIR hyperspectral imaging combined with image processing, Computers and Electronics in Agriculture 103 (2014) 1e10. G. El-Masry, D.W. Sun, P. Allen, Chemical-free assessment and mapping of major constituents in beef using hyperspectral imaging, Journal of Food Engineering 117 (2013) 235e246. H. Pu, D.W. Sun, J. Ma, D. Liu, M. Kamruzzaman, Hierarchical variable selection for predicting chemical constituents in lamb meats using hyperspectral imaging, Journal of Food Engineering 143 (2014) 44e52.

466 SECTION j III Application fields [10] A. Iqbal, D.W. Sun, P. Allen, An overview on principle, techniques and application of hyperspectral imaging with special reference to ham quality evaluation and control, Food Control 46 (2014) 242e254. [11] M. Vidal, J.M. Amigo, Pre-processing of hyperspectral imaging: a review of best practice, performance and pitfalls for in-line and on-line applications, Journal of Near Infrared Spectroscopy 20 (2012) 483e508. [12] F. Mertens, S. Pa¨tzold, G. Welp, Spatial heterogeneity of soil properties and its mapping with apparent electrical conductivity, Journal of Plant Nutrition and Soil Science 171 (2008) 146e154. [13] N. Robinson, P. Rampant, A. Callinan, M. Rab, P. Fisher, Advances in precision agriculture in south-eastern Australia. II. Spatio-temporal prediction of crop yield using terrain derivatives and proximally sensed data, Crop & Pasture Science 60 (2009) 859e869. [14] X. Jin, L. Kumar, Z. Li, H. Feng, X. Xu, G. Yang, J. Wang, A review of data assimilation of remote sensing and crop models, European Journal of Agronomy 92 (2018) 141e152. [15] G.P. Petropoulos, K. Arvanitis, N. Sigrimis, Hyperion hyperspectral imagery analysis combined with machine learning classifiers for land use/cover mapping, Expert Systems with Applications 39 (3) (2012) 3800e3809. [16] A. Kross, H. McNairn, D. Lapen, M. Sunohara, C. Champagne, Assessment of RapidEye vegetation indices for estimation of leaf area index and biomass in corn and soybean crops, International Journal of Applied Earth Observation and Geoinformation 34 (2015) 235e248. [17] H. Li, Z.X. Chen, Z.W. Jiang, W.B. Wu, J.Q. Ren, B. Liu, T. Hasi, Comparative analysis of GF-1, HJ-1, and Landsat-8 data for estimating the leaf area index of winter wheat, Journal of Integrative Agriculture 16 (2) (2017) 266e285. [18] J.G.P.W. Clevers, L. Kooistra, M.M.M. van den Brande, Using Sentinel-2 data for retrieving LAI and Leaf and canopy chlorophyll content of a potato crop, Remote Sensing 9 (5) (2017) 405. [19] A. Laliberte, A. Rango, A. Slaughter, Unmanned aerial vehicle (UAVs) for rangeland remote sensing, in: Proceedings of the 3rd Annual Symposium Research Insights in Semiarid Ecosystems. Tucson, Arizona, U.S.A, 2006. [20] A. Apan, A. Held, S. Phinn, J. Markley, Detecting sugarcane ‘orange rust’disease using EO-1 hyperion hyperspectral imagery, International Journal of Remote Sensing 25 (2004) 489e498. [21] A. Del Fiore, M. Reverberi, A. Ricelli, F. Pinzari, S. Serranti, A. Fabbri, G. Bonifazi, C. Fanelli, Early detection of toxigenic fungi on maize by hyperspectral imaging analysis, International Journal of Food Microbiology 144 (2010) 64e71. [22] P.P.J. Roosjen, B. Brede, J.M. Suomalainen, H.M. Barthoolomeus, L. Kooistra, J.G.P.W. Clevers, Improved estimation of leaf area index and leaf chlorophyll content of a potato crop using multi-angle spectral data e potential of unmanned aerial vehicle imagery, International Journal of Applied Earth Observation and Geoinformation 66 (2018) 14e26. [23] Y. Zhong, A. Ma, Y.S. Ong, Z. Zhu, L. Zhang, Computational intelligence in optical remote sensing image processing, Applied Soft Computing 64 (2018) 75e93. [24] P. Raghavendra, T. Pullaiah, Advances in Cell and Molecular Diagnostics, Academic Press, London, United Kingdom, 2018. [25] G.P. Asner, R.E. Martin, Spectranomics: emerging science and conservation opportunities at the interface of biodiversity and remote sensing, Global Ecology and Conservation 8 (2016) 212e219.

Hyperspectral imaging in crop fields: precision agriculture Chapter j 3.3

467

[26] D.C. Marvin, L.P. Koh, A.J. Lynam, S. Wich, A.B. Davies, R. Krisnamurthy, E. Stokes, R. Starkey, G.P. Asner, Integrating technologies for scalable ecology and conservation, Global Ecology and Conservation 7 (2016) 262e275. [27] V. Proenc¸a, L.J. Martin, H.M. Pereira, M. Ferna´ndez, L. McRae, J. Belnap, M. Bohm, N. Brummitt, J. Garcı´a-Moreno, R.D. Gregory, J.P. Honrado, N. Jurgens, M. Opige, D.S. Schmeller, P. Tiago, C.A.M. Van Swaay, Global biodiversity monitoring: from data sources to essential biodiversity variables, Biological Conservation 213 (2017) 256e263. [28] A. Verma, R. van der Wal, A. Fischer, Imaging wildlife: new technologies and animal censuses, maps and museums, Geoforum 75 (2016) 75e86. [29] A. Plaza, J.A. Benediktsson, J. Boardman, J. Brazile, L. Bruzzone, G. Camps-Valls, J. Chanussot, M. Fauvel, P. Gamba, J. Gualteri, M. Marconcini, J.C. Tilton, G. Trianni, Recent advances in techniques for hyperspectral image processing, Remote Sensing of Environment 113 (2009) 110e122. [30] S. Kalluri, Z. Zhang, J. Jaja, S. Liang, J. Townshend, Characterizing land surface anisotropy from AVHRR data at a global scale using high performance computing, International Journal of Remote Sensing 22 (2001) 2171e2191. [31] A. Plaza, C.I. Chang, High Performance Computing in Remote Sensing, Taylor and Francis, Boca Rato´n, Florida, U.S.A, 2007. [32] J. Dorband, J. Palencia, U. Ranawake, Commodity computing clusters at goddard space flight center, Journal of Space Communications 3 (2003) 1. [33] A. Lastovetsky, J. Dongarra, High-Performance Heterogeneous Computing, Wiley., New York, New York, U.S.A, 2009. [34] A. Plaza, D. Valencia, J. Plaza, P. Martı´nez, Commodity cluster based parallel processing of hyperspectral imagery, Journal of Parallel and Distributed Computing 66 (3) (2006) 345e358. [35] A. Plaza, J. Plaza, D. Valencia, Impact of platform heterogeneity on the design of parallel algorithms for morphological processing of high-dimensional image data, The Journal of Supercomputing 40 (1) (2007) 81e107. [36] A. Plaza, D. Valencia, J. Plaza, An experimental comparison of parallel algorithms for hyperspectral analysis using homogeneous and heterogeneous networks of workstations, Parallel Computing 34 (2) (2008) 92e114. [37] S. Hauck, The roles of FPGAs in reprogrammable systems, Proceedings of the IEEE 86 (1998) 615e638. [38] E. Lindholm, J. Nickolls, S. Oberman, J. Montrym, NVIDIA Tesla: a unified graphics and computing architecture, IEEE Micro 28 (2008) 39e55. [39] W. Jetz, J. Cavender-Bares, R. Pavlick, D. Schimel, F.W. Davis, G.P. Asner, R. Guralnick, J. Kattge, A.M. Latimer, P. Moorcroft, M.E. Schaepman, M.P. Schildhauer, F.D. Schneider, F. Schrodt, U. Stahl, S.L. Ustin, Monitoring plant functional diversity from space, Nature Plants 2 (2016) 16024. [40] W. Turner, Sensing biodiversity, Science 346 (6207) (2014) 301e302. [41] S.E. Sesnie, P.E. Gessler, B. Finegan, S. Thessler, Integrating Landsat TM and SRTM-DEM derived variables with decision trees for habitat classification and change detection in complex neotropical environments, Remote Sensing of Environment 112 (2008) 2145e2159. [42] C.A. Lee, S.D. Gasster, A. Plaza, C.I. Chang, B. Huang, Recent developments in high performance computing for remote sensing: a review, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 4 (2011) 508e527.

468 SECTION j III Application fields [42a] C. Nansen, L.D. Geremias, Y. Xue, F. Huang, J.R.P. Parra. Agricultural case studies of classification accuracy, spectral resolution, and model over-fitting, Applied Spectroscopy 67 (11) (2013) 1332e1338. [43] J.M. Amigo, H. Babamoradi, S. Elcoroaristizabal, Hyperspectral image analysis. A tutorial, Analytica Chimica Acta 896 (2015) 34e51. [44] C.I. Chang, X. Jiao, C.C. Wu, Y. Du, M.L. Chang, A review of unsupervised spectral target analysis for hyperspectral imagery, EURASIP Journal on Applied Signal Processing 2010 (2010) 503752. [45] D.M. Haaland, H.D.T. Jones, M.H. Van Benthem, M.B. Sinclair, D.K. Melgaard, C.L. Stork, M.C. Pedroso, P. Liu, A.R. Brasier, N.I. Andrews, D.S. Lidke, Hyperspectral confocal fluorescence imaging: exploring alternative multivariate curve resolution approaches, Applied Spectroscopy 63 (2009) 271e279. [46] S. Sayad, Real Time Data Mining, Self-Help Publishers, Cambridge, Ontario, Canada, 2011. [47] S. Vanonckelen, S. Lhermitte, A. Van Rompaey, The effect of atmospheric and topographic correction methods on land cover classification accuracy, International Journal of Applied Earth Observation and Geoinformation 24 (2013) 9e21. [48] Y.B. Cai, H. Zhang, W.B. Pan, Y.H. Chen, X.R. Wang, Land use pattern, socio-economic development, and assessment of their impacts on ecosystem service value: study on natural wetlands distribution area (NWDA) in Fuzhou City, southeastern China, Environmental Monitoring and Assessment 185 (2013) 5111e5123. [49] J. Sturck, A. Poortinga, P.H. Verburg, Mapping ecosystem services: the supply and demand of flood regulation services in Europe, Ecological Indicators 38 (2014) 198e211. [50] T. Blaschke, G.J. Hay, M. Kelly, S. Lang, P. Hofmann, E. Addink, R. Queiroz-Feitosa, F. Van Der Meer, H. Van Der Werff, F. Van Coillie, D. Tiede, Geographic object-based image analysis - towards a new paradigm, ISPRS Journal of Photogrammetry and Remote Sensing 87 (2014) 180e191. [51] A. Ma, Y. Zhong, B. Zhao, H. Jiao, L. Zhang, Semi supervised subspace- based DNA encoding and matching classifier for hyperspectral remote sensing imagery, IEEE Transactions on Geoscience and Remote Sensing 54 (8) (2016) 4402e4418. [52] W. He, H. Zhang, L. Zhang, H. Shen, Hyperspectral image denoising via noise-adjusted iterative low-rank matrix approximation, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 8 (6) (2015) 3050e3061. [53] X. Tong, X. Xu, A. Plaza, H. Xie, H. Pan, W. Cao, D. Lv, A new genetic method for subpixel mapping using hyperspectral images, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 99 (2016) 1e12. [54] L. Yue, H. Shen, J. Li, Q. Yuan, H. Zhang, L. Zhang, Image super-resolution: the techniques, applications, and future, Signal Processing 128 (2016) 389e408. [55] A. Gupta, Y.S. Ong, L. Feng, Multifactorial evolution: towards evolutionary multitasking, IEEE Transactions on Evolutionary Computation 20 (3) (2016) 343e357. [56] W. He, H. Zhang, L. Zhang, Sparsity-regularized robust non-negative matrix factorization for hyperspectral unmixing, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 9 (9) (2016) 4267e4279. [57] X. Xu, X. Tong, A. Plaza, Y. Zhong, H. Xie, L. Zhang, Joint sparse sub-pixel mapping model with endmember variability for remotely sensed imagery, Remote Sensing 9 (1) (2016) 1e20. [58] S. Asseng, F. Ewert, C. Rosenzweig, J.W. Jones, J.L. Hatfield, A. Ruane, K.J. Boote, P. Thorburn, R.P. Ro¨tter, D. Cammarano, N. Brisson, B. Basso, P. Martre, P.K. Aggarwal,

Hyperspectral imaging in crop fields: precision agriculture Chapter j 3.3

[59]

[60]

[61]

[62]

[63]

[64]

[65]

[66]

[67]

[68]

[69] [70]

469

C. Angulo, P. Bertuzzi, C. Biernath, J. Doltra, S. Gayler, R. Goldberg, R. Grant, L. Heng, J.E. Hooker, L.A. Hunt, J. Ingwersen, R.C. Izaurralde, K.C. Keserbaum, C. Mu¨ller, S. Naresh Kumar, C. Nendel, G. O’ Leary, J.E. Olesen, T.M. Osborne, T. Palosuo, E. Priesack, D. Ripoche, M.A. Semenov, I. Shcherbak, P. Steduto, C.O. Sto¨ckle, P. Stratonovitch, T. Streck, I. Supit, M. Travasso, F. Tao, K. Waha, D. Wallach, J.W. White, J. Wolf, Uncertainty in simulating wheat yields under climate change, Nature Climate Change 3 (2013) 827e832. S. Bassu, N. Brisson, J.L. Durand, K. Boote, J. Lizaso, J.W. Jones, C. Rosenzweig, A.C. Ruane, M. Adam, C. Baron, B. Basso, C. Biernath, H. Boogaard, S. Coinjin, M. Corbeels, D. Deryng, G. De Sanctis, S. Gayler, P. Grassini, J. Hatfield, S. Hoek, C. Izaurralde, R. Jongschaap, A.R. Kemanian, K.C. Kersebaum, S.H. Kim, N.S. Kumar, D. Makowski, C. Mu¨ller, C. Nendel, E. Priesack, M.V. Pravia, F. Sau, I. Shcherbak, F. Tao, E. Teixeira, D. Timlin, K. Waha, How do various maize crop models vary in their responses to climate change factors? Global Change Biology 20 (2014) 2301e2320. F. Ewert, R.P. Ro¨tter, M. Bindi, H. Webber, M. Tmka, K.C. Kersebaum, J.E. Olesen, M.K. Van Ittersum, S. Janssen, M. Rivington, M.A. Semenov, D. Wallach, J.R. Porter, D. Stewart, J. Verhagen, T. Gaiser, T. Paloosuo, F. Tao, C. Nendel, P.P. Roggero, L. Bartosova, S. Asseng, Crop modelling for integrated assessment of risk to food production from climate change, Environmental Modelling & Software 72 (2015) 287e303. H.C.J. Godfray, J.R. Beddington, I.R. Crute, L. Haddad, D. Lawrence, J.F. Muir, J. Pretty, S. Robinson, S.M. Thomas, C. Toulmin, The Challenge of Food Security Science, Edward Elgar Publishing Co., Cheltenham, United Kingdom, 2010. A. Kern, Z. Barcza, H. Marjanovic, T. Arendas, N. Fodor, P. Bonis, P. Bognar, J. Lichtenerger, Statistical modelling of crop yield in central Europe using climate data and remote sensing vegetation indices, Agricultural and Forest Meteorology 260 (2018) 300e320. E.S. Mohamed, A.M. Ali, M.A. El-Shirbeny, A. El-Razek, A. Afaf, I.Y. Savin, Nearinfrared spectroscopy techniques for soil contamination assessment in the Nile Delta, European Journal of Soil Science 49 (6) (2016) 632e639. K. Tan, Y. Ye, Q. Cao, P. Du, J. Dong, Estimation of arsenic contamination in reclaimed agricultural soils using reflectance spectroscopy and ANFIS model, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing 7 (2014) 2540e2546. E. Choe, K.W. Kim, S. Bang, I.H. Yon, K.Y. Lee, Quantitative analysis and mapping of heavy metals in an abandoned Au-Ag mine area using NIR spectroscopy, Environmental Geology 58 (2009) 477e482. R.F. Kokaly, R.N. Clark, Spectroscopic determination of leaf biochemistry using banddepth analysis of absorption feature and stepwise multiple linear regression, Remote Sensing of Environment 167 (1999) 267e287. G. Siebielec, G. McCarty, T.I. Stuczynski, J.B. Reeves, Near and mid-infrared diffuse reflectance spectroscopy for measuring soil metal content, Journal of Environmental Quality 33 (2004) 2056e2069. S.C. Dunagan, M.S. Gilmore, J.C. Varekamp, Effects of mercury on visible/near-infrared reflectance spectra of mustard spinach plants, Environment and Pollution 148 (1) (2007) 301e311. Y. Wu, X. Zhang, Q. Liao, J. Ji, Can contaminant elements in soils be assessed by remote sensing technology: a case study with simulated data? Soil Science 176 (2011) 196e205. L.Y. Yang, X.H. Gao, W. Zhang, F.F. Shi, L.H. He, W. Jia, Estimating heavy metal concentrations in topsoil from vegetation reflectance spectra of hyperion images: a case study of Yushu County, Qinghai, China, Journal of Applied Ecology 27 (6) (2016) 1775e1784.

470 SECTION j III Application fields [71] M. Liu, X. Liu, M. Li, M. Fang, W. Chi, Neural-network model for estimating leaf chlorophyll concentration in rice under stress from heavy metals using four spectral indices, Biosystems Engineering 106 (2010) 223e233. [72] C. Zhou, D. Wang, S. Chen, Y. Liu, M. Wang, Vegetation corrected depth model and its application in mineral extraction from HSI, Journal of China University of Geosciences 40 (2015) 1365e1370. [73] T. Kemper, S. Sommer, Use of airborne hyperspectral data to estimate residual heavy metal contamination and acidification potential in the Guadiamar flood plain Andalusia, Spain after the Aznacollar mining accident, Proceedings of SPIE Remote Sensing for Environmental Monitoring, GIS Applications, and Geology 5574 (2004) 224e234. [73a] E. Choe, F. van der Meer, F. van Ruitenbeek, H. van der Werff, B. de Smeth, K.-W. Kima. Mapping of heavy metal pollution in stream sediments using combined geochemistry, field spectroscopy, and hyperspectral remote sensing: a case study of the Rodalquilar mining area, SE Spain, Remote Sensing of Environment 112 (2008) 3222e3233. [74] C.M. Pandit, G.M. Filipelli, L. Li, Estimation of heavy metal contamination in soil using reflectance spectroscopy and partial least-square regression, International Journal of Remote Sensing 31 (2010) 4111e4123. [75] A. Riaza, E. Garcı´a-Mele´ndez, A. Mueller, Spectral identification of pyrite mud weathering products: a field and laboratory evaluation, International Journal of Remote Sensing 32 (2011) 185e208. [76] P.H. Rathod, C. Brackhage, F.D. Van der Meer, I. Mu¨ller, M.F. Noomen, D.G. Rossiter, G.E. Dudel, Hyperion data for estimation of forest leaf area index, IEEE Transactions on Geoscience and Remote Sensing 41 (4) (2017) 916e921. [77] E. Husson, F. Lindgren, F. Ecke, Assessing biomass and metal contents in riparian vegetation along a pollution gradient using an unmanned aircraft system, Water, Air, and Soil Pollution 225 (6) (2014) 1e14. [78] M. Li, Based on the spectral variation of vegetation monitoring the heavy metal pollution of soil, Earth Science Frontiers 16 (2009) 86e87. [79] X. Li, J. Li, X. Liu, Collaborative inversion heavy metal stress in rice by using twodimensional spectral feature space based on HJ-1 A HIS and radarsat-2 SAR remote sensing data, International Journal of Applied Earth Observation and Geoinformation 78 (2019) 39e52. [80] J. Tuominen, T. Lipping, Detection of Environmental Change Using Hyperspectral Remote Sensing, Working Report 2011-26, Posiva Inc, Eurajoki, Finland, 2011. [81] P. Ceccato, S. Flasse, S. Tarantola, S. Jacquemoud, J.M. Gregoire, Detecting vegetation water content using reflectance in the optical domain, Remote Sensing of Environment 77 (2001) 22e33. [82] C.L. Jones, P.R. Weckler, N.O. Maness, M.L. Stone, R. Jayasekara, Estimating water stress in plant using hyperspectral sensing, in: Annual International Meeting Sponsored by ASAE, 2004. Paper number 043065. [83] E.S. Koksal, Y. Gungor, Y.E. Yildirim, Spectral reflectance characteristics of sugar beet under different levels of irrigation water and relationships between growth parameters and spectral indexes, Irrigation and Drainage 60 (2010) 187e195. [84] Q. Yi, A. Bao, Q. Wang, J. Zhao, Estimation of leaf water content in cotton by means of hyperspectral indices, Computers and Electronics in Agriculture 90 (2013) 144e151. [85] A.B. Gonza´lez-Ferna´ndez, J.R. Rodrı´guez-Pe´rez, V. Marcelo, Using field spectrometry and a plant probe accessory to determine leaf water content in comercial vineyards, Agricultural Water Management 156 (2015) 43e50.

Hyperspectral imaging in crop fields: precision agriculture Chapter j 3.3

471

[86] J. Borzuchowski, K. Schulz, Retrieval of leaf area index and soil water content using hyperspectral remote sensing under controlled glass house conditions for spring barley and sugar beet, Remote Sensing 2 (7) (2010) 1702e1721. [87] S. Elsayed, Comparing the performance of active and passive reflectance sensors to assess the normalized relative canopy temperature and grain yield of drought-stressed barley, Field Crops Research 177 (2015) 148e160. [88] K.K. Bandyopadhyay, S. Pradhan, R.N. Sahoo, R. Singh, V.K. Gupta, Characterization of water stress and prediction of yield of wheat using spectral indices under varied water and nitrogen management, Agricultural Water Management 146 (2014) 115e123. [89] Y. Kim, D.M. Glenn, J. Park, H.K. Ngugi, B.L. Lehman, Hyperspectral image analysis for plant stress detection, in: American Society of Agricultural and Biological Engineers, Annual International Meeting, Pittsburgh, Pennsylvania, 2010. [90] Y. Sadeh, X. Zhu, K. Chenu, D. Dunkerley, Sowing date detection at the field scale using CubeSats remote sensing, Computers and Electronics in Agriculture 157 (2019) 568e580. [91] M. Landi, B. Giovanni, Protecting crop species from biotic and abiotic constraints in the era of global change: are we ready for this challenge? American Journal of Agricultural and Biological Sciences 11 (2016) 51e53. [92] J.S. West, C. Bravo, R. Oberti, D. Lemaire, D. Moshou, H.A. McCartney, The potential of optical canopy measurement for targeted control of field crop diseases, Annual Review of Phytopathology 41 (2003) 593e614. [93] C. Nansen, N. Elliott, Remote sensing and reflectance profiling in entomology, Annual Review of Entomology 61 (2016) 139e158. [94] J.R. Riley, Remote sensing in entomology, Annual Review of Entomology 34 (1) (1989) 247e271. [95] P.I. Kerchev, B. Fenton, C.H. Foyer, R.D. Hancock, Plant responses to insect herbivory: interactions between photosynthesis, reactive oxygen species and hormonal signalling pathways, Plant, Cell and Environment 35 (2) (2012) 441e453. [96] P.D. Nabity, J.A. Zavala, E.H. DeLucia, Indirect suppression of photosynthesis on individual leaves by arthropod herbivory, Annals of Botany 103 (4) (2008) 655e663. [97] J.T. Trumble, D.M. Kolodny-Hirsch, I.P. Ting, Plant compensation for arthropod herbivory, Annual Review of Entomology 38 (1) (1993) 93e119. [98] W.E. Riedell, T.M. Blackmer, Leaf reflectance spectra of cereal aphid-damaged wheat, Crop Science 39 (6) (1999) 1835e1840. [99] N. Elliott, M. Mirik, Z. Yang, T. Dvorak, M. Rao, J. Michels, T. Walker, V. Catana, M. Phoofolo, T. Royer, Airborne multi-spectral remote sensing of Russian wheat aphid injury to wheat, Southwestern Entomologist 32 (4) (2007) 213e219. [100] D. Reisig, L. Godfrey, Remote sensing for detection of cotton aphide(homoptera: aphididae) and spider mitee(acari: tetranychidae) infested cotton in the San Joaquin Valley, Environmental Entomology 35 (6) (2006) 1635e1646. [101] C.R. Silva, A.E. Olthoff, J.A.D. de la Mata, A.P. Alonso, Remote monitoring of forest insect defoliation. A review, Forest Systems (3) (2013) 377e391. [102] G.F. Backoulou, N.C. Elliott, K. Giles, M. Phoofolo, V. Catana, Development of a method using multispectral imagery and spatial pattern metrics to quantify stress to wheat fields caused by Diuraphis noxia, Computers and Electronics in Agriculture 75 (1) (2011) 64e70. [103] G.F. Backoulou, N.C. Elliott, K. Giles, M. Phoofolo, V. Catana, M. Mirik, J. Michels, Spatially discriminating Russian wheat aphid induced plant stress from other wheat stressing factors, Computers and Electronics in Agriculture 78 (2) (2011) 123e129.

472 SECTION j III Application fields [104] G.F. Backoulou, N.C. Elliott, K.L. Giles, M.N. Rao, Differentiating stress to wheat fields induced by Diuraphis noxia from other stress causing factors, Computers and Electronics in Agriculture 90 (2013) 47e53. [105] R. Baker, C. Sansford, C. Jarvis, R. Cannon, A. MacLeod, K. Walters, The role of climatic mapping in predicting the potential geographical distribution of non-indigenous pests under current and future climates, Agriculture, Ecosystems & Environment 82 (2000) 57e71. [106] O. Mutanga, A. Skidmore, Red edge shift and biochemical content in grass canopies, ISPRS Journal of Photogrammetry and Remote Sensing 62 (2007) 34e42. [107] J. Behmann, J. Steinru¨cken, L. Plu¨mer, Detection of early plant stress responses in hyperspectral images, ISPRS Journal of Photogrammetry and Remote Sensing 93 (2014) 98e111. [108] F. Martinelli, R. Scalenghe, S. Davino, S. Panno, G. Scuderi, P. Ruisi, P. Villa, D. Stroppiana, M. Boschetti, L.R. Goulart, Advanced methods of plant disease detection. A review, Agronomy for Sustainable Development 35 (2015) 1e25. [109] S. Thessler, L. Kooistra, F. Teye, H. Huitu, A.K. Bregt, Geosensors to support crop production: current applications and user requirements, Sensors 11 (2011) 6656e6684. [110] P. Mehl, Y.R. Chen, M.S. Kim, D.E. Chen, Development of hyperspectral imaging techniques for the detection of apple defects and contaminants, Journal of Food Engineering 61 (2004) 67e81. [111] G.A. Leiva-Valenzuela, R. Lu, J.M. Aguilera, Prediction of firmness and soluble solids content on blueberries using hyperspectral reflectance imaging, Journal of Food Engineering 115 (2013) 91e98. [112] J. Qin, R. Lu, Detection of pits in tart cherries by hyperspectral transmission imaging, Transactions of the American Society of Agricultural Engineers 48 (2005) 1963e1970. [113] A. Baiano, C. Terracone, G. Peri, R. Romaniello, Application of hyperspectral imaging for prediction of physico-chemical and sensory characteristics of table grapes, Computers and Electronics in Agriculture 87 (2012) 142e151. [114] J. Zhao, Q. Ouyang, Q. Chen, J. Wang, Detection of bruise on pear by hyperspectral imaging sensor with different classification algorithms, Sensor Letters 8 (2010) 570e576. [115] G. El-Masry, N. Wang, A. ElSayed, M. Ngadi, Hyperspectral imaging for non-destructive determination of some quality attributes for strawberry, Journal of Food Engineering 81 (2007) 98e107. [116] R.P. Cogdill, C.R. Hurburgh, G.R. Rippke, Single-kernel maize analysis by near-infrared hyperspectral imaging, Transactions of the American Society of Agricultural Engineers 47 (2004) 311e320. [117] H. Zhang, J. Paliwal, D.S. Jayas, N.D.G. White, Classification of fungal infected wheat kernels using near-infrared reflectance hyperspectral imaging and support vector machine, Transactions of the American Society of Agricultural and Biological Engineers 50 (2007) 1779e1785. [118] Y.L. Liu, Y.R. Chen, C.Y. Wang, D.E. Chan, M.S. Kim, Development of a simple algorithm for the detection of chilling injury in cucumbers from visible/near-infrared hyperspectral imaging, Applied Spectroscopy 59 (2005) 78e85. [119] M. Taghizadeh, A.A. Gowen, C.P. O’Donnell, Comparison of hyperspectral imaging with conventional RGB imaging for quality evaluation of agaricus bisporus mushrooms, Biosystems Engineering 108 (2011) 191e194.

Hyperspectral imaging in crop fields: precision agriculture Chapter j 3.3

473

[120] B.K. Cho, M.S. Kim, I.S. Baek, D.Y. Kim, W.H. Lee, J. Kim, H. Bae, Y.S. Kim, Detection of cuticle defects on cherry tomatoes using hyperspectral fluorescence imagery, Postharvest Biology and Technology 76 (2013) 40e49. [121] J. Xing, P. Hung, S. Symons, M. Shahin, D. Hatcher, Using a short wavelength infrared (SWIR) hyperspectral imaging system to predict alpha amylase activity in individual canadian western wheat kernels, Sensing and Instrumentation for Food Quality and Safety 3 (2009) 211e218. [122] N.N.D. Trong, M. Tsuta, B.M. Nicolai, J. De Baerdemaeker, W. Saeys, Prediction of optimal cooking time for boiled potatoes by hyperspectral imaging, Journal of Food Engineering 105 (2011) 617e624. [123] D. Lorente, N. Aleixos, J. Gomez-Sanchis, S. Cubero, J. Blasco, Selection of optimal wavelength features for decay detection in citrus fruit using the ROC curve and neural networks, Food and Bioprocess Technology 6 (2013) 530e541.