Deep neural networks (NNs) trained on hyperspectral images are employed typically for the classification of new images collected from the same sensor, assuming similar characteristics to those of the training images. Creating, however, high-quality ground truth (GT) for training is rather complex, especially when attempting to classify multi-temporal images over seasonal changes. To overcome this difficulty, we propose a novel method that utilizes an additional, one-time collection of hyperspectral FENIX images in the Spring along with ground observations from the end of the Fall. The hyperspectral data are then used for simulation of GT for training. At the same time, the field campaign allows for fine-tuning of the NN to achieve enhanced, multi-seasonal hyperspectral image classification. Indeed, we demonstrate how the proposed method successfully classifies new VEN μS images obtained during different seasons.

Training a deep neural network for classification constitutes a major problem in remote sensing due to the lack of adequate field data. Acquiring high-resolution ground truth (GT) by human interpretation is both cost-ineffective and inconsistent. We propose, instead, to utilize high-resolution, hyperspectral images for solving this problem, by unmixing these images to obtain reliable GT for training a deep network. Specifically, we simulate GT from high-resolution, hyperspectral FENIX images, and use it for training a convolutional neural network (CNN) for pixel-based classification. We show how the model can be transferred successfully to classify new mid-resolution VENμS imagery.

Among forest types, the Mediterranean maquis is specifically exposed to fluctuations in water availability. Therefore, monitoring the water-use patterns of its major tree species is key in quantifying the local and regional water balance. However, the traditional measurement methods of tree water-use at high spatial scales are difficult and labor-intensive, thus indirect methods become useful. Evaporation of water from the stomatal pores on the leaf surface involves evaporative cooling, and hence the differences between the leaf temperature and its surrounding air temperature (Delta T-l(eaf-air)) can serve as a reliable estimator for tree water-use. Here, we used direct measurement of transpiration rate (E) with a gas exchange system, simultaneously with ground Thermal Infra-Red (TIR) imaging to study the relationship between Delta T-l(eaf-air) and E in dominant tree species of the Mediterranean maquis. Controlled experiments were conducted in parallel with measurements in the forest, on five tree species of contrasting leaf shapes (Conifers: Pines halepensis; Cupressus sempervirens; Broadleaf: simple: Quercus calliprinos; Ceratonia siliqua; compound: Pistacia lentiscus). Next, we used a quantitative approach, applying a leaf energy balance model to estimate E from the TIR images and compared it to the direct measurement of the gas exchange system. We report evaporative cooling across the five species, replicated in tree saplings and in mature trees in the forest. The conifers were significantly cooler than broadleaves by up to similar to 4 degrees C and produced narrower Delta T-l(eaf-air) ranges. Estimations of E from Delta T-l(eaf-air) were relatively close to the observed E, with some overestimations. Our observations show that TIR imaging can detect transpiration-related differences in Delta T-l(eaf-air) among species and can be used to estimate E in natural environments. Yet, the dependence of Delta T-l(eaf-air) on E is species-specific and thus, empirical associations must be developed separately for each of the species.

The production of sweet basil (Ocimum basilicum L.) in Israel during the winter exposes the crop to temperatures below 12 degrees C, which promote visible chilling injuries, reducing yields by up to 85%. In 2011, the addition of illumination was examined as an agro-technique to mitigate visible chilling injuries under a semi-commercial operation. Lamps were maintained at 30100 cm above plants' tops, and agronomical, morphological and physiological parameters were evaluated. Artificial illumination increased the mean and minimum temperatures between 7.1 and 1.1 degrees C, and 2.4 and 2.2 degrees C compared to out-door or non-illuminated tunnels, respectively. Leaf temperatures were increased by 35 degrees C. Illuminated plants yielded more (P < 0.05) than non-illuminated plants, regardless of lamp height. Highest (P < 0.05) yields were recorded when lamps were situated 4090 cm above plants' tops. The effect of illumination on yield was related to increased leaf area and internodes length. Morning dew was reduced by 85% in plants subjected to artificial illumination; and visible chilling injuries were alleviated by 8085%, regardless of lamp height. The results indicate that increased leaf and air temperature is likely the fundamental chilling injury mitigating factor. Interestingly, leakage of soluble and accumulation of antioxidant, phenols and rosmarinic acid was similar in illuminated and non-illuminated plants. This discrepancy between the horticultural and the physiological parameters should be studied further, as artificial illumination might have additional mechanisms affecting chilling injury other than direct temperature increase. (C) 2016 Elsevier GmbH. All rights reserved.

To use remotely sensed spectral data for determining rates and timing of variable rate nitrogen (N) applications at a commercial scale, the most reliable indicators of crop N status must be determined. This study evaluated the ability of hyperspectral remote sensing to predict N stress in potatoes (Solanum tuberosum) during two growing seasons (2010 and 2011). Spectral data were evaluated using ground based measurements of leaf N concentration. Two canopy-scale hyperspectral images were acquired with an AISA-Eagle hyperspectral camera in both years. The experiment included five N treatments with varying rates and timing of N fertilizer and two potato cultivars, Russet Burbank (RB) and Alpine Russet (AR). Partial Least Squares regression (PLS) models resulted in the best prediction of leaf N concentration (r2=0.79, Root Mean Square Error of Cross Validation (RMSECV)=14% across dates for RB; r2=0.77, RMSECV=13% across dates for AR). Applying the Nitrogen Sufficiency Index (NSI) formula to spectral indices/models made them mostly insensitive to the effects of cultivar. The most promising technique for determining N stress in potato based on spectral indices was found to be the MERIS Terrestrial Chlorophyll Index (MTCI) due to a combination of relatively high r2 values, lower RMSECVs, and high accuracy assessment. Pairwise comparison tests from the means separation showed that spectral indices/models from the imagery resulted in more statistically significant groupings of crop stress levels for the spectra than leaf N concentration because canopy-scale spectral data are affected by both tissue N concentration and biomass. The results of this study suggest that upon proper sensor calibration, canopy-scale spectral data may be the most sensitive tool available to detect N status of a potato crop.

The nitrogen sufficiency index (NSI) can be used for in-season variable rate management of nitrogen (N) fertilizer to maintain productivity of potato (Solanum tuberosum, L.) while reducing leaching losses. The objective of this study was to evaluate the implications of using high spatial resolution broad-band imagery for determining N prescriptions at different growth stages. Aerial images were obtained for research plots, as well as for a commercial potato field (59 ha) near Becker, Minnesota on 30, 56 and 79 days after emergence (DAE) with a Redlake MS4100 multispectral camera. In research plots, experimental treatments included five N treatments with varying rates and timing of N fertilizer, and two potato varieties, Russet Burbank and Alpine Russet. Spectral indices investigated in this study adequately predicted N stress based on leaf N concentration (r (2) values within dates ranged from 0.49 to 0.82). On 56 and 79 DAE, the Green Ratio Vegetation Index (GRVI) normalized by an NSI that used the recommended rate and timing from the research plots as a reference showed that most areas of the commercial field did not require supplemental N fertilizer (using an NSI over-sufficiency threshold of 120 %). Based on regional guidelines, N was over-applied to the commercial field, but in situations where N is applied more sparingly, a GRVI NSI threshold of 80 % should be used to identify areas that are most suitable for supplemental N fertilizer. A practical approach and the implications associated with using spectral data for in-season N management are proposed.

Potato yield and quality are highly dependent on an adequate supply of water. In this study, 3 years of information from thermal and RGB images were collected to evaluate water status in potato fields. Irrigation experiments were conducted in commercial potato fields (Desiree; drippers). Two water-deficit scenarios were tested: a short-term water deficit (by suppressing irrigation for a number of days before image acquisition), and a long-term cumulative water deficit. Ground and aerial images were acquired in various phenological stages along the potato growing season. Effects of irrigation treatments were recorded by thermal indices and biophysical measurements of stomatal conductance (SC), leaf water potential, leaf osmotic potential and gravimetric water potential in soil. Canopy temperature was delineated from the thermal images with and without fused information from the RGB image. Crop water stress index (CWSI) was calculated, using three forms of minimum baseline temperature: empirical, theoretical and statistical. An empirical evaluation of maximum baseline temperature of Tair + 7 °C was used in all CWSI forms examined. Statistical tests and comparison of CWSI with biophysical measurements were performed to evaluate the responses to irrigation treatments. The results indicated a high correlation of CWSI with SC from tuber initiation to maturity based on ground and aerial data (0.64 ≤ R2 ≤ 0.99). Similar trends of increasing CWSI from well to deficit-irrigated treatments were found in all three growing seasons. The results also showed that CWSI may be calculated based merely on thermal imagery data.

Hyperspectral images and spectroradiometer measurements were taken from cauliflower (Brassica oleracea, Botrytis group), aubergine (Solanum melongena) and kohlrabi (Brassica oleracea, Gongylodes group) plants in a controlled experiment. Plants were grown in media with sodium chloride (NaCl) concentrations between 30 and 150 mmol. Spectral and spatial processing techniques were developed to assess the ability to distinguish between plants exposed to various levels of salinity stress. Local autocorrelation analysis was used to detect the spatial patterns that characterise the effects of salinity on crop canopy. This analysis was applied on a vegetation index in the spectral range of 435-554 nm, the green indigo ratio (GIR) index. The processing strategies that were developed were able to distinguish three levels of salinity effects. The strategy based on a combined spatial-spectral index yielded the most consistent results with average total accuracy of 62%, whereas accuracies obtained with known spectral vegetation indices were 29%. The presented method may be implemented in other cases of vegetation stresses where symptoms are characterised by patchiness and can be imaged, not necessarily in the visible spectral range (400-750 nm). (C) 2012 lAgrE. Published by Elsevier Ltd. All rights reserved.

Potato yield and quality are highly dependent on an adequate supply of water. Thermal Infrared (TIR) imagery data can be used for calculating the Crop Water Stress Index (CWSI), which may be used as a means for managing the dynamic demands of water in potato fields during the growth season. Field experiments were conducted in 2010, 2011 and 2012 in commercial potato fields (Solanum tuberosum L. cv. Desiree; drip irrigation) at Kibbutz Ruhama, Israel (31.38° N, 34.59° E). Two scenarios of water deficit were tested, a short term water deficit, induced by suppressing irrigation for a number of days before field campaign, and a long term cumulative water deficit induced by reduced seasonal application. TIR and RGB images were used for delineation of canopy temperature and calculation of CWSI throughout the season. Plant water status was evaluated by measuring leaf stomatal conductance (SC). Data from 2011 showed that there is a statistically significant effect (α = 0.05) of the cumulative irrigation treatments on tuber yield, and that tuber yield is highly correlated with CWSI ((0.62≤R2≤0.93). These findings may imply the potential of using CWSI values as a threshold for managing irrigation, and will be further examined with data from the 2012 spring season.

Spectroradiometric measurements were taken from eggplant (Solanum melongena) and cauliflower (Brassica oleracea) plants in a controlled experiment treated with sodium chloride (NaCl) concentrations between 30 and 150 mmol. Assessment of seven vegetation indices indicated potential difference between three levels of salinity in effects, with the green indigo ratio (GIR of reflectance of 554 and 436 nm) yielding the most consistent results. Salinity effects started from the second week, but the highest differences were obtained in the fourth week for cauliflower and in the third and sixth weeks for eggplant.

Efficient real-time discrimination of image objects is greatly affected by their radiometry, which is only partly accounted for by image scene calibration. Such calibration treats mainly variations in flux density in the generalized imaged scene plane rather than on the objects’ surface. The proposed methodology uses ratios between secondary parameterizations: e.g., absorption features and spectral derivatives. Clustering in the ratios’ parameter space may allow differentiation between image objects despite limitations regarding their relative calibration. The usefulness of this approach was demonstrated in the challenging task of separating Mediterranean vegetation species using imaging spectroscopy.