Vegetation parameters are important for the global carbon cycle. Therefore, the quantitative acquisition of vegetation parameters is crucial. The inverse process of the PROSAIL model has provided a classic method for vegetation parameter estimation. The current PROSAIL inverse process simulation, based on lookup tables and other methods, has challenges, such as low accuracy and poor spatial universality. To address these issues, this study proposed a PROSAIL inverse process simulation method based on deep learning. The PROSAIL model was decomposed according to the physical process of the model. Then, the corresponding network module was designed based on the inverse process of each module and combined into the IPROSAIL-Net. This network uses the reflectance of the vegetation canopies as the input and presents the leaf structure parameter, chlorophyll a + b content Cab, equivalent water thickness, dry matter content Cm, and leaf area index (LAI) as the output. This article conducted experiments using two different sets of data. PROSAIL simulation data were used to invert the five values. The inversion accuracy was above 0.99 when the number of training samples reached more than 30000. EnMAP data were used to invert the Cab and LAI values. When the number of training samples reached more than 120, the cereals accuracy was above 0.97 and the maize and rapeseed accuracies were above 0.99. The IPROSAIL-Net design was further split into six networks for separate trainings to verify its rationality. Therefore, the IPROSAIL-Net neural network is reasonable and feasible for remote sensing inversion of vegetation parameters.

The accuracy of a classifier, whether it is an ensemble or not, is directly influenced by the training data used in learning. In remote sensing, training data mislabeling is inevitable and faces a major challenge. This paper proposes a versatile data cleaning which handles the mislabeling problem by exploiting the ensemble concepts for identifying, then eliminating or correcting the mislabeled training data. A powerful ensemble method, random forest, is at the core of our filter design and helps to distinguish mislabeled data from uncorrupted data more accurately. The major contribution of this work lies on the explicit use of the hypothesis margin as a decision means to identify and eliminate or correct mislabeled training data in an ensemble learning framework. Another key development that makes our algorithm superior to existing approaches is a design that avoids rare class instances to be mistaken for class noise. This fundamental aspect makes our data cleaning system particularly suitable for remote sensing classification tasks which usually suffer from both mislabeling and imbalance problems. The effectiveness of our algorithm is demonstrated in performing mapping of land covers. The generalization performance of two major supervised noise-sensitive classifiers, boosting and K-nearest neighbors, is strengthened by effective class noise reduction. A comparative analysis is conducted with respect to random forest, deep convolutional neural networks, as well as two well-established ensemble-based class noise filters, the majority vote and the consensus vote filters. This analysis demonstrates that our approach is more accurate than deep convolutional neural networks (one-dimensional CNN, AlexNet, EfficientNet, ResNet50 and ShuffletNet) and the reference ensemble methods.

Remote sensing of leaf chlorophyll content (LCC) is critical in precision agriculture, forest monitoring, and pest management. Chlorophyll absorbs light in the visible and red-edge spectrum areas. However, many multispectral sensors lack red-edge bands. Subsequently, the development of effective visible chlorophyll-related indicators is important. The hue angle defined in the CIE xy chromaticity diagram, which reflects the human eye’s color discrimination ability relying on visible light (400–700 nm), is introduced in this study for LCC estimation. This study proposes a new multispectral hue angle calculation method employing D65 standard illuminant data, investigates the effect of illuminant factors and band settings on hue angle calculation, and presents tristimulus weight coefficients for calculating MERIS, MSI, and ETM+ hue angles. Furthermore, the sensitivity of the link between LCC and different hue angles, as well as vegetative indices (VIs), is assessed in dense and sparse canopies. The results show that using wavelength-independent illuminant data or discrete bands overcalculates the reflectance-based hue angle. Moreover, our proposed hue angle calculated using D65 standard illuminant data is more sensitive to LCC than the existing hue angle using wavelength-independent illuminant data. In dense canopies, the multispectral hue angle has a stronger relationship with LCC than in sparse canopies. The multispectral hue angle generally appears weaker or similar to the red-edge VIs (MTCI and CIred-edge), but stronger than the non-red-edge VIs (CIgreen, NDVI, NGRDI, and TGI). These findings imply that the multispectral hue angle calculated from visible bands provides the potential to monitor LCC of large-scale plants.

Quantitative estimation of regional leaf area index (LAI) is an important basis for large-scale crop growth monitoring and yield estimation. With the development of deep learning, theoretically, the use of neural networks can effectively improve the accuracy of LAI estimation, but sufficient training samples are often required due to a large number of network parameters. In an actual regional LAI quantitative estimation, there are only a few samples, which is difficult to train in networks. Therefore, a crop dual-learning generative adversarial network (CROP-DualGAN) was proposed in this article for data enhancement of small samples to estimate regional LAI. The method uses dual learning to generate hyperspectral reflectance and corresponding LAI, including two groups of generative adversarial networks, in which the generator is used to generate data that conforms to the distribution of the training set, and the discriminator is used to judge the true or false generated samples. The generators and discriminators are constantly optimized in the confrontation so that the distribution of generated data is closer to that of training samples. In single crop type experiments, 30 training samples with enhanced in VGG16 achieved the R2of cereal, maize and rape seed as 0.921, 0.990 and 0.956, and in SSLLAI-Net achieved the R2of cereal, maize and rape seed as 0.971, 0.991 and 0.962. In multiple crop types experiments, the result is lower than individual crop estimation, but higher than that of without enhancement. Finally, non-parametric test is used to prove that most improvement in LAI estimation is significant, and the accuracy won’t decrease when improvement is not significant. In all, proposed method is universal and can effectively help benchmark models to improve regional LAI estimation accuracy with neural networks.

This paper presents the results of a collaboration between Italian and Chinese research groups carried out under the context of the GEO work programme and AfricultuReS H2020 project. The paper encompasses three main aspects: (a) the description of the results achieved on the high resolution crop mapping carried out in some African countries in the framework of the AfricultuReS project; (b) the description of the crop early warning service delivered in the framework of the AfricultuReS project; and (c) the application of the desert locust disaster monitoring model to the case of Somali. Using a multi-source data approach, the factors that have an important influence on the desert locust occurrence and spread process were extracted. The connection between the three points mentioned above must be sought in the fact that the combination of an accurate mapping of agricultural areas accompanied by techniques for estimating any threats to them allows to accurately estimate the possible effect in terms of production loss and food security.

Spectral–spatial-based deep learning models have recently proven to be effective in hyper-spectral image (HSI) classification for various earth monitoring applications such as land cover classification and agricultural monitoring. However, due to the nature of “black-box” model representation, how to explain and interpret the learning process and the model decision, especially for vegetation classification, remains an open challenge. This study proposes a novel interpretable deep learning model—a biologically interpretable two-stage deep neural network (BIT-DNN), by incorporating the prior-knowledge (i.e., biophysical and biochemical attributes and their hierarchical structures of target entities)-based spectral–spatial feature transformation into the proposed framework, capable of achieving both high accuracy and interpretability on HSI-based classification tasks. The proposed model introduces a two-stage feature learning process: in the first stage, an enhanced interpretable feature block extracts the low-level spectral features associated with the biophysical and biochemical attributes of target entities; and in the second stage, an interpretable capsule block extracts and encapsulates the high-level joint spectral–spatial features representing the hierarchical structure of biophysical and biochemical attributes of these target entities, which provides the model an improved performance on classification and intrinsic interpretability with reduced computational complexity. We have tested and evaluated the model using four real HSI data sets for four separate tasks (i.e., plant species classification, land cover classification, urban scene recognition, and crop disease recognition tasks). The proposed model has been compared with five state-of-the-art deep learning models. The results demonstrate that the proposed model has competitive advantages in terms of both classification accuracy and model interpretability, especially for vegetation classification.

The presence of noise is often unavoidable and has been a serious nuisance factor that needs to be taken into account in the hyperspectral image classification. Effective noise handling is one of the most difficult problems in data classification. Ensemble-based filtering has been demonstrated successful in dealing with the class noise problem. In this paper, a novel two-step ensemble-based data filtering method is proposed to improve the hyperspectral image classification accuracy in the presence of class noise. The proposed method is a combination of noise redundancy classifiers and sensitive algorithms. The experimental results on two public hyperspectral datasets demonstrate the effectiveness of the proposed approach.

In this paper, we propose a novel ensemble margin based semi-supervised random forest (EMRF) algorithm for the classification of the hyperspectral image with limited training data. The proposed method tries to improve the effectiveness of the ensemble model via adaptively labeling the unlabeled instances with high classification probability then adding them into the training set. The classification probability of a training instance is reflected by the unsupervised margin value of this instance. The higher ensemble margin of an instance, the higher probability the instance being classified correctly and added into to the training set in the next iteration.

The Leaf are index (LAI), a key factor in many physiological processes in plants, plays a crucial role in characterizing vegetation. Various vegetation indices (VIs) have been proposed to estimate such an important parameter. Some issues, however, have been found such as low accuracy and poor robustness. In this study, a Modified Triangular Difference Vegetation Index (MTDVI) was proposed by selecting optimal spectral bands on our previously reported TDVI. It was estimated by retrieving the LAI of summer corn and winter wheat. The results showed that the MTDVI was more sensitive than TDVI and some commonly used VIs to the LAIs of summer corn. To show the performance, another MTDVI0 was used to compared the MTDVI, indicating that MTDVI0 had better coefficient of determination (R2) but larger root mean square error (RMSE). The study shows that the proposed MTDVI is satisfactory to retrieve the LAI of wheat and corn, and can be also considered for other similar crops.

Tree pests can cause rapid and widespread damage, reducing the economic value of plants, production, in the case of fruit trees, and their role in mitigating climate change. There are several diseases that affect trees, including, for example, pine tree nematode (PWN), trunk fungal diseases, or Xylella fastidiosa (Xf).Mapping of diseased plants based on visual or automatic analysis of remote sensing data could be a useful support for in situ investigation planning. However, there is a clear need for better modeling methods to elaborate potential critical scenarios in order to early detect diseases (e.g. Xf) in host plants.Maxent (Maximum Entropy) has proved powerful when modeling species with available scarce presence-only occurrence data. The purpose is to predict potential distributions or explore expanding distributions. In this work we applied the Maxent model comparing local modeling results with worldwide cases towards a more comprehensive analysis of potential pest risk zones.

Monitoring and mapping weeds within agricultural crops is required for the implementation of precision agriculture approaches such as patch spraying. A precise and targeted weed control would bring about positive consequences from both environmental and economic perspectives. Given the small spectral differences between crop species, VNIR hyperspectral data can be a powerful tool to perform an effective weed monitoring and identification when high spatial and spectral resolution data is available (i.e. UAV platforms). This work explores the spectral differences between crops and weeds to evaluate the ability of UAV hyperspectral data to separate maize crop from weeds and to discriminate different types of weeds. To this aim, UAV and field hyperspectral data were acquired in some maize fields in Italy during the 2016 growing season. Results showed that by exploiting leaf chlorophyll and carotenoid contents, retrieved using spectral indices or by inverting PROSAIL, is it possible to discriminate between maize crop and weeds and, moreover, among weed types. The procedure allowed the quantification of crop/weeds relative ground cover, which showed a good relationship with the corresponding measured relative LAI values.

Rotation forest (RoF) is a powerful ensemble classifier and has been demonstrated the outstanding performance in hyperspectral data classification. However, the classification task suffers from the class imbalanced problem which has been considered to be one of the most important challenges. The traditional construction method of RoF biases classifying the majority classes and ignores recognizing the minority classes samples. This letter proposes a novel adaptive ensemble method based on SMOTE and RoF with differentiated sampling rates (AdaSRoF) for the multiclass imbalance problem. The proposed method adaptively generates several balanced data sets with more diversity and less noise by using SMOTE and a dynamic data sampling ratio for base classifiers. The obtained results on two publicly available hyperspectral images show that the proposed method can get more diversity and better performance than support vector machine (SVM), random forest (RF), and RoF in multiclass imbalance learning.

In this paper, we propose a novel Synthetic Minority Oversampling Technique based Rotation forest (SMOTERoF) algorithm for the classification of imbalanced hyperspectral image data. The main idea of the proposed method is to iteratively balance the class distribution of training set by SMOTE for each rotation decision tree. Experiment results on the hyperspectral image Indian Pines AVRIS with different imbalance ratio (IR) show that our algorithm obtains better classification performance compared with Rotation Forest (RoF), random undersampling, random oversampling, SMOTE, as well as Under sampling based RoF (UnderRoF) which is an extended version of UnderBagging.

Growth stage information is a very primary parameter for managing the grain crop. In this study, time series Enhanced Vegetation Index (EVI) data in 2015 were obtained in Anhui Province, China based on 8-day composite MOD09A1 data products, and the extraction method of rice phenology was specifically analyzed at a provincial scale. HANTS (Harmonic Analysis of Time Series) filtering algorithm was firstly used to smooth the time series EVI curves and identify every growth and development period of paddy rice. The results indicate that HANTS has a better performance in removing noise, restoring original information and retain the change information of multi-temporal EVI curves. The primary planting regions were located in the southern Huai-River areas, from early June to late June in 2015. It turned into the transplanting stage in succession from south to north and followed by the heading stage in succession in early August. From early September to late October, it turned into mature stage from south to north in succession.

This paper analyzes four practical change detection algorithms with Landsat-8 data on landslide mapping in the Kaikoura earthquake happened on November 14, 2016. Eleven band groups built from seven reflective bands of Landsat-8 data were used in the experiment. Total 21 change detection results based on various combinations of band groups and algorithms were obtained. The results were qualitatively and quantitatively analyzed based on manual interpretation and the ROC curve. It shows that all results have high false alarm rates and the accuracy varies for different combinations. Further analysis indicates that false alarms are those being easily affected by phenology factors and human activities, such as cropland, river, snow, and shadow etc. The high false alarm rate indicated by the ROC curve is additionally due to small but hard to avoid errors in the preparation of ground truths. Furthermore, change vector analysis offers plenty of information but does not directly enhance landslide. Based on above analysis, several possible ways to improve the change detection based landslide monitoring method are given at the end of the paper.

The paper describes the preliminary results of the January-August 2015 multi-frequency EO data acquisition campaign conducted over the Maccarese (Central Italy) farm. From January to May radar Cosmo SkyMed Ping-Pong (HH-VV), RapidEye and ZY-3 multispectral VHR optical images, as well as in situ data, have been acquired to retrieve biophysical and/or bio-chemical characteristics of soil and crops. LAI trend has been analyzed and compared by using both polarimetric and optical retrieval algorithms while soil moisture measurements have been compared with the radar backscattering.

A multispectral full-waveform light detection and ranging (LiDAR) instrument prototype with four wavelengths and a supercontinuum laser as a light source was designed to monitor the fine structure and the biochemical parameters of vegetation. Components of the instrument included a 2-D scanning platform, a supercontinuum laser source, a receiving optical system, and a multichannel full-waveform measurement module. The LiDAR instrument can simultaneously measure multichannel-returned full-waveform laser signals. Position information in the recorded waveform allowed us to compute the distance from the target, whereas the intensity of the signal provided the spectral reflectance. Performance for the measuring distance and the spectrum was evaluated. Experiments indicated that the instrument has high measurement accuracy and has the ability to detect the biochemical characteristics of vegetation via construction of the normalized difference vegetation index and the photochemical reflectance index. The experiment also indicated that the instrument has the potential to generate spectral 3-D point clouds. Therefore, the instrument could play a significant role in detecting the vertical distribution of structural and biochemical characteristics of vegetation.

Wheat aphid, Sitobion avenae F. is main aphid species infesting winter wheat in the filling stage in Northwest China, and it has severe impact on both wheat yield and quality. The study acquired hyperspectral data by ASD FieldSpec Pro spectrometer at the canopy level and aphid damage levels of samples in the filling stage of winter wheat. The spectral characteristics of wheat uninfected by aphid and healthy wheat were analyzed, then the correlation simulating analysis model (CSAM) which was established by a 2-dimensional coordinate system with average spectral of healthy wheat samples called also base spectrum as abscissa axis and the spectral of other samples as vertical axis respectively is developed and tried to monitor the aphid damage levels. It is concluded that the fitting curves obtained by the reflectance of samples relative to healthy wheat samples are near to straight line in the range from 400nm to 1000nm (R2>0.99), and the slopes of fitting lines decrease as aphid damage levels become serious. Moreover, the most sensitive band regions were selected out. The result shows that the correlation between the slopes of fitting line and aphid damage levels is the highest in the range from 400 nm to 810 nm (R2=0.89). Therefore, the CSAM can be sued to discriminate the aphid damage levels in the filling stage of winter wheat.

Wheat aphid, Sitobion avenae F. is the most destructive insect infesting winter wheat and appears almost annually in northwest China. Past studies have demonstrated the potential of remote sensing for detecting diseases and insects damage. In the study, hyperspectral imaging in the visible and near-infrared (500-900nm) region was tried to determinate aphid of wheat leaf and detect damage region of winter leaf caused by aphid. The principal component analysis (PCA) and spectral indices which used to monitor some stresses were applied to extract aphid information. The result showed that the classification result was better based on the second principal component (PC2) image and the third principal component (PC3) image by principal component (PC) transformation than spectral indices. Then, the mean reflectance of pixels with aphid and pixels without aphid was obtained, respectively, and the most sensitive reflectance regions to aphid were selected in visible and near-infrared by comparing the reflectance difference of two classes. Further, Leaf aphid damage index (LADI) was established according to two the sensitive reflectance region, and the leaf region with aphid, the infested leaf region and healthy leaf region were classified by LADI value of image. The result showed that the aphid damage area ratio of each wheat leaf estimated by pixels number of three classes was consistent with the survey the damage area ratio. So LADI had potential for detecting the leaf damage region caused by aphid.

Leaf area index (LAI) of crop canopies is important for crop growth monitoring and yield estimation. Considering the practical need of achieving distribution properties of LAI at a special spatial scale, and the difficult acquisition of corresponding observations at the same scale, a method integrating multi-resolution data at larger scales based on continuous wavelet theory is proposed to provide a more effective LAI dataset. For this method, firstly multi-scale wavelet theory is selected for multi-resolution data decomposition, and then decomposed signals and statistics of observations are coupled for wavelet reconstruction. Finally, the new constructed data is used for LAI estimation through multiple linear regression method. Barley is selected as experimental object. The performance of this method is quantitatively analyzed by testing indicators, i.e. Number of effective bands, R2, and MRA. Theory analysis and numerical practices fully confirm the feasibility and validity of the proposed method in crop LAI estimation.

Drought is a main meteorological disaster leading to significant environmental, social and economic consequences due to its high frequency and large influence. Traditional methods of drought monitoring are to simplify drought into drought indices derived from weather station data. With the development of remote sensing, It is necessary to assess the potential of such indices derived from a specific sensor for drought monitoring in a particular region before we apply this remote-sensed measure in large area. This study assesses the SPIs in different time scales and their performance in identifying drought, as well as the potential of MODIS-derived spectral indices in drought monitoring in Hebei, an arid and semi-arid areas of Northern China. Reveal the relationship between different MODIS data and meteorological SPI in different time scales, so as to identify their feasibility for drought monitoring.

In traditional, the research of wheat stripe rust spreading mode was focus on the large scale meteorological propagation model. In this paper, we present a diagnosis and propagation model of wheat stripe rust in field scales based on Bayesian network, which can provides technical support for short-term accurate prediction and precision pesticide of stripe rust in small scale. Through the wheat stripe rust induction experiment in different scales field grids, we get the spatio-temporal occurring and spreading law of wheat stripe rust in different scales. Then we can choice the accurate scale of model, structure learning method and parameters learning method of the spatio-temporal spreading model of wheat stripe rust. After constructing the Bayesian network model, we can predict and analyze incidence of stripe rust in a certain fields and can verify the accuracy of model by contrast the real disease in field with the predict result.

Stripe Rust, caused by Puccinia striiformis f. sp. tritici, has emerged as one of the most destructive foliar diseases of winter wheat (Triticum aestivum L.) and caused great yield loss. The primary objective of this study is to identify the stress characteristics of wheat leaves caused by such a disease. After performing artificial inoculation of stripe rust, hundreds of leaves with different infections were collected and leaf reflectance were conducted using ASD's (Analytical Spectral Devices) leaf clip probe. Based on the nonequivalent optical properties on the opposite side of a leaf, their spectral characteristics were specifically compared and the correlations between disease severity indices and water or chlorophyll were further analyzed. The final results indicated that, for the face and back sides of stripe rust infected wheat leaves, some obvious differences existed in the spectral reflectance and correlation curves between foliar disease severity indexes (FDSIs) and reflectance, especially with the increase of infections.

Aphid (Hemiptera: Aphididae) outbreaks appear in wheat (Triticum aestivum L.) planting area in China, and had significant economic impacts on wheat. It has severe impact on both winter wheat yield and grain quality. The aim of this study was to monitor and predict of wheat aphid by analyzing the relationship between land surface temperature (LST), modified normalized difference water index (MNDWI) and the occurrence and prevalence of aphid in wheat. The results showed that LST was an important driving factor for occurrence of aphid, and MNDWI was sensitive to aphid damage degree. Meanwhile, a 2-dimensional feature space was established based on LST and MNDWI derived from Landsat TM images, and discrimination model of aphid damage degrees was established according to the distribution of samples in the feature space. It was verified that the overall accuracy of discrimination model is 82.5%, and kappa accuracy is 73.88%.

Based on the theory of radiation transfer model, this paper modified the Simultaneous Heat and Water model to calculate FPAR vertical distribution in maize canopy and analyzed the relationships between FPAR and some parameters like maize canopy structure, solar zenith, soil reflectance, etc. The validation results using field measurements prove the model to be accurate.

Aphid (Hemiptera: Aphididae) appears in wheat planting area of China almost every year and have had significant economic impacts on wheat yield. As a result, large amounts of insecticides are used to control aphid populations, which may cause environmental pollution. Therefore, remote sensing as a repeatable and rapid method is necessary for monitoring aphid damage level. The study analyzed the hyperspectral characteristics of wheat infested by several aphid damage levels and selected out the sensitive bands to aphid damage levels, and the aphid damage hyperspectral index (ADHI) was developed based on the most sensitive bands to aphid damage levels in the visible, near-infrared and short-wave infrared regions. The results indicated that ADHI exhibited a high correlation with aphid damage levels (R2=0.839), so it had the potential to detect wheat damage caused by aphid.

Land surface temperature (LST) is a key parameter in numerous environmental studies. In order to lower the subpixel temperature estimation error caused by re-sampling of remote sensing data, a disaggregation method for subpixel temperature using the remote sensing endmember index based technique (DisEMI) was established in this study. To take an advantage of simultaneous, multi-resolution observations at coincident nadirs by the Advanced Spaceborne Thermal Emission Reflection Radiometer (ASTER) and the MODerate-resolution Imaging Spectroradiometer (MODIS), LST products from the two sensors were examined for a portion of suburb area in Beijing, China. The verified results indicate that the estimated temperature distribution was basically consistent with that of ASTER LST product(R2= 0.709 and RMSE = 2.702 K).

This paper presents a new forest leaf area index (LAI) inversion method from multisource and multiangle data combined with a radiative transfer model and the strategy of -means clustering and artificial neural network (ANN). Four scenes of Landsat-5 Thematic Mapper (L5TM) and Beijing-1 small satellite multispectral sensors (BJ1) images, acquired at different times, were selected to construct multisource and multiangle image data in this study. Considering a vertical distribution of forest LAI from both overstory and understory, a hybrid model of the invertible forest reflectance model (INFORM) was used to support the retrieval of forest LAI to eliminate the dependence of understory vegetation. The simulated data from INFORM outputs, added with a random noise, were first clustered by -means method, and were then trained by ANN to obtain the inversion model for each group (cluster). Next, the inversion model was applied to the different combinations of multiangle data to retrieve the forest LAI. Finally, a validation of inverted results with Moderate Resolution Imaging Spectroradiometer LAI product and field measurements was conducted. The experimental results indicate that the accuracy of the inverted forest LAI can be improved through the addition of observation angle data, if the quality of the image data is ensured. The inversion accuracy of LAI with the multiangle image data is improved by 30% compared to the average accuracy of the inverted LAI with the single angle data after considering the addition of random noise to the ANN training data.

Among the multisource data fusing methods, the potential advantages of remote sensing of solar-reflective visible and near-Infrared [(VNIR); 400-900 nm] data and thermal-infrared (TIR) data have not been fully mined. Usually, a linear unmixed method is used for the purpose, which results in low estimation accuracy of subpixel land-surface temperature (LST). In this paper, we propose a novel method to estimate subpixel LST. This approach uses the characteristics of high spatial-resolution advanced spaceborne thermal emission and reflection radiometer (ASTER) VNIR data and the low spatial-resolution TIR data simulated from ASTER temperature product to generate the high spatial-resolution temperature data at a subpixel scale. First, the land-surface parameters (e.g., leaf area index, normalized difference vegetation index (NDVI), soil water content index, and reflectance) were extracted from VNIR data and field measurements. Then, the extracted high resolution of land-surface parameters and the LST were simulated into coarse resolutions. Second, the genetic algorithm and self-organizing feature map artificial neural network (ANN) was utilized to create relationships between land-surface parameters and the corresponding LSTs separately for different land-cover types at coarse spatial-resolution scales. Finally, the ANN-trained relationships were applied in the estimation of subpixel temperatures (at high spatial resolution) from high spatial-resolution land-surface parameters. The two sets of data with different spatial resolutions were simulated using an aggregate resampling algorithm. Experimental results indicate that the accuracy with our method to estimate land-surface subpixel temperature is significantly higher than that with a traditional method that uses the NDVI as an input parameter, and the average error of subpixel temperature is decreased by 2-3 K with our method. This method is a simple and convenient approach to estimate subpixel LST from high spatial-temporal resolution data quickly and effectively.

Satellite image simulation is one of the key methods to check the expected performance of the satellites before they launched or when satellites can not provide images in other time. In order to provide a useful tool to analyze whether the payload of HJ-1B (a small satellite of the environment-monitoring constellation) is enough, we develop a simulation system for the infrared cameras, which consists of four bands including NIR band (0.75-1.10¿m), SWIR (1.55-1.75¿m), MIR (3.50-3.90¿m) and TIR (10.5-12.5¿m). The spatial resolution of NIR and SWIR band is 150 meter, while 300 meter for the MIR and TIR band. The sensor is an optical-mechanics multi-scanning system with maximum scanning degree of 29 degree.

The effect of crop leaf angle on the canopy-reflected spectrum cannot be ignored in the inversion of leaf area index (LAI) and the monitoring of the crop-growth condition using remote-sensing technology. In this paper, experiments on winter wheat (Triticum aestivumL.) were conducted to identify the crop leaf angle distribution (LAD) by two-temporal (erecting and elongation stages) and bidirectionalin situreflected spectrum and the Airborne Multiangle Thermal Infrared (TIR) Visible Near-Infrared (VNIR) Imaging System (AMTIS) images. The distribution characters of the leaf angle for different LAD varieties were expressed using the beta-distribution function and the SAILTH radiative transfer models. The proportion of the leaf angle in 5deg angle classes (from 5deg to 90deg) for erectophile, planophile, and horizontal varieties was dominated by 75deg, 55deg, and 35deg. The different LAD varieties had a similar canopy reflectance in 680 nm (red) and 800 nm (near-infrared band) at the erecting stage, while they had significant differences at the elongation stage. The ratio of the canopy reflectance of 800 nm at the erecting stage [R800(B)] to the canopy reflectance of 800 nm at the elongation stage [R800(A)] was used to identify the different LAD varieties through the selected two-temporal canopy reflectance. A method based on the semiempirical model of the bidirectional reflectance distribution function (BRDF) was also introduced in this paper. The structural parameter-sensitive index (SPEI) was used in this paper for crop LAD identification. SPEI is proved to be more sensitive to identify erectophile, planophile, and horizontal LAD varieties than the structural scattering index and the normalized difference f-index. We found that it is feasible to identify horizontal, planophile, and erectophile LAD varieties of wheat by studying two-temporal and bidirectional canopy-reflected spectrum

A novel instrument to measure photochemical reflectance index (PRI) and nitrogen reflectance index (NRI) is presented. It can measure the incident and reflected radiance of vegetation at 531 nm, 570 nm and 670 nm bands by the 6 specially designed interference filters, then PRI and NRI is derived. The PRI index can be used for the diagnosis of crop stress, such as waterlogging and drought. The NRI can be used for the diagnosis of crop nitrogen nutrition deficiency. The instruments can diagnose the plant growth status by the acquired spectral response. This optical instrument includes photoelectric detector module, signal process and A/D convert module, the data storing and transmission module and human-machine interface module. The detector is the core of the instrument which measures the spectrums at special bands. The microprocessor calculates the PRI and NRI value based on the A/D value. And the value can be displayed on the instrument's LCD, stored in the flash memory of instrument and can also be uploaded to PC through the PC's RS232 serial interface. The prototype was tested in the crop field at different view directions. The PRI and NRI data acquired by the novel instrument are compared and calibrated by an ASD FieldSpec spectrometer, which shows that this instrument is successfully developed and reliable. The winter wheat's leaf area index, water content, chlorophyll density and nitrogen content are successfully detected by the PRI and NRI instrument, this novel instrument is portable and could be fixed on the agricultural machine traveling in the field, which demonstrates the promising application of the novel portable instrument.

This work focused on developing the appropriate spectral index to monitor yellow rust disease, which is relatively insensitive to species, canopy structures (such as leaf area index and leaf distribution angles), foliar inner structures, and soil condition variations. Firstly, canopy spectra and canopy chlorophyll concentration were acquired during growth duration, and the relationship between SPAD value and measured chlorophyll concentration was established. It indicated that the logarithmic relationship exists between TCARI/OSAVI and SPAD value, with a coefficient of determination R/sup 2/ = 0.7795 (n = 320). Therefore, the combined index (CCII) can be used to minimize the effects of LAI and no photosynthetic materials on the retrieval of chlorophyll concentration at canopy level. Secondly, The photochemical reflectance index (PRI) was chosen for disease index inversion by minimizing internal leaf structure effect, has and the linear negative relationship exists between the normalized photochemical reflectance index (NPRI) and disease index (DI), with a coefficient of determination R/sup 2/ = 0.8477 (n = 63), so the normalized photochemical reflectance index (NPRI) can be used to monitor the disease index of yellow rust.

The objective in this study was to develop proper vegetation indices for prediction of soil irrigation demanding under vegetation covering conditions. The traditional method for the winter wheat yellow rust field survey is time consuming. It was discussed of the selection method of characteristic spectral bands and the establishing of inversion model to monitor winter wheat yellow rust using hyperspectral data in this study. The correlation coefficients between selected vegetation index and disease incidence (DI) at infected stages. Inversion models between DI and vegetation index such as normalized difference vegetation index (NDVI), ratio vegetation index (RVI), transformed vegetation index (TVI) were used to monitor yellow rust. The multi-temporal hyperspectral airborne images were acquired from winter booting stage to milking stage, and the yellow rust disease of winter wheat was analyzed using hyperspectral images. Compared with healthy wheat, spectral reflectance of disease wheat was higher in 560-670 nm bands but lower in near infrared bands and the absorption depth of chlorophyll in red band and reflectance peak in green band are relatively reduced. A novel spectral index for yellow rust indices was presented, and the degree and area of yellow rust disease were successfully remotely sensed from the multi-temporal hyperspectral data based on spectral index.

Remote sensing offers the potential to determine rapidly the physiological condition of crop over large areas. Canopy reflectance data selected at key growth stages of winter wheat were analyzed. It indicated that winter wheat growth phonological changes can be evaluated by red edge variables (REV) characteristics. Regression equations between chlorophyll concentration, nitrogen concentration, soluble sugar were established, prediction of foliar soluble sugar and chlorophyll concentration by red edge position, foliar nitrogen concentration by the amplitude of red edge, starch concentration by the near infrared plateau, leaf area index by the area of red edge peak are successful and feasible. For this purpose red edge variables can play a vital role in providing time-specific and time-critical information for precision farming, due to their capabilities in measuring canopy spectrum variability.

The objective in this study was to develop the time-specific and time-critical method to overcome the limitations of traditional field sampling methods for variable rate fertilization. Farmers, agricultural managers and grain processing enterprises and interested in measuring and assessing soil and crop status in order to apply adequate fertilizers quantities to crop growth. This work focused on studying the relationship between SPAD readings and crop chlorophyll concentration and/or nitrogen content to determine the amount of nitrogen fertilizer recommended for variable rate management in precision agriculture. The recommended rate of conventional uniform rate fertilizer management was higher than that of variable management. The grain yield, ear numbers, 1000-grain weight and grain protein content were measured between CK and variable rate fertilizer treatments. It indicated that variable rate fertilization reduced the variability of wheat yield, ear numbers and 1000-grain weight, but it didn't increased crop yield and grain protein content significantly, compared to traditional nitrogen application. The nitrogen fertilizer use efficiency was improved, for this purpose, the variable rate technology (VRT) based on SPAD readings could be used to prevent the water pollution and environmental deterioration.

The seedtime plays an important role on the growth, yield and quality of winter wheat. Firstly, the seedtime was successfully remotely sensed by NDVI derived from a LandSat TM image captured in elongation stage. Secondly, an optimized yield estimation model was designed based on NDVI and seedtime, and the optimized model was successfully tested by 3 LandSat TM images captured in heading, grain filling and milking stages.

A method for estimating grain quality of winter wheat using Landsat Thematic Mapper data and GIS data is presented, and it gave the development and application of the Geographical Information Application system for estimating wheat grain quality. The system, supported by COMGIS, JModel-Base Management System and other advanced Spatial Information Technologies, can deal with the fusion analysis of remotely sensed data, GIS data and other multisources data, flexible management of models and knowledge

There is considerable interest in assessing leaf area index (LAI) to evaluate crop growth and production. In this article, an operational approach was proposed to evaluating LAI of summer maize for different cultivars under different nitrogen treatments and developmental stages by selecting ten familiar remote sensing vegetation indices (VIs). The result indicated that VIs had the potential for faithfully estimating LAI, and the estimation power of VIs for assessing LAI was best from bell stage to silking stage, which primarily depended on the LAI dynamic variation during the process of the growth. When the estimation power of VIs was systematically verified with the other independent data set, the VIs could accurately evaluate LAI. The ratio spectral index of R/sub 810//R/sub 560/ was the best index to estimate LAI, which was wondrously sensitive to LAI dynamic variation almost without the influence of cultivars, growth stages and nitrogen treatments. The exponential regression model for LAI based on R/sub 810//R/sub 560/ was also established with a mean determination of coefficient (R/sup 2/) of 0.9573 (P value=0.01) and a mean root mean square error (RAISE) of 0.0365. Therefore, the spectral index of R/sub 810//R/sub 560/ could be considered as a sensitive indicator as LAI of summer maize.

The objective in this study was to develop proper vegetation indices for prediction of soil irrigation demanding under vegetation covering conditions. The four tested winter wheat varieties were under four different irrigation treatment levels at two different growth stages, one at booting stage and another at grain filling stage, the winter wheat canopy and bare soil reflected spectrum, foliar and soil moisture condition were tested in the field at an interval of 7 days from erecting stage to ripening stage. The combined spectral indices of photochemical reflectance index (PRI) and normalized difference water index (NDWI) were chosen in this study. Regression equations between the ration of NPRI to NDWI and soil water content (SWC) were found to be most appropriate for assessing soil water condition under vegetation covering, and relationship between the ration of PRI to NDWI and soil irrigation demanding conditions were also studied. It indicated that soil irrigation demanding condition under vegetation covering conditions could be evaluated by a pragmatic approach for canopy level estimation with proper combined vegetation index

In this study, the selection method of characteristic spectral band and the establishing of inversion model to monitor winter wheat stripe rust using hyperspectral data is discussed. The correlation coefficients between the DI (disease incidence) at different stages of infection and the initial canopy reflectance spectral and the derivative of the reflectance spectrum were compared, respectively. The results showed that the derivative of the reflectance spectra has reached higher significant level with the DI than the initial reflectance spectral data. The initial reflectance in the visible light 680 nm wavelength and the near infrared 976 nm, 1010 nm wavelength were selected to do regression with the DI of winter wheat stripe rust. And some inversion models between the DI and the hyperspectral data or its conversion patterns like NDVI (Normalized difference vegetation index), RVI (Ratio vegetation index), TVI (Transformed vegetation index) and its differential values of the canopy spectral reflectance data to monitor winter wheat stripe rust were established. Meanwhile, those correlation coefficients were compared respectively, of which we found the pattern of vegetation index has more efficient commonly than initial canopy spectral reflectance data by aggression analysis with the DI. The paper also suggested that the possibility of developing a special visible/near-infrared sensor for the detection of the DI of winter wheat stripe rust theoretically. Else, the SRSI (stripe rust stress index) mechanism model was presented for the first time in This work.

The wheat grain quality and its influence were introduced, and the mechanism and methods to forecast grain quality were studied. Firstly, the leaf nitrogen content at anthesis stage was proved to be an indicator of grain protein content, and the spectral indices significant correlated to leaf nitrogen content at anthesis stage were the potential indictors for grain protein content. The vegetation index. VIgreen, derived from the spectral reflectance at green and red bands, was significant correlated to the leaf nitrogen content at anthesis stage, and also high significant to the final grain protein content. Secondly, the environment stress, such as irrigation, fertilization, temperature, had important influence on grain quality. The irrigation stress can increase grain protein content. The leaf water content depends on irrigation levels, therefore, the spectral indices correlated to leaf water content were also potential indictors for grain quality. The spectral reflectance at SWIR band and other water indices at seeding filling stage were proved to be high significant correlate to grain protein. Finally, the predicting models of grain protein content were built based on the transfer principle of leaf nitrogen content and the effect of irrigation stress

This paper mainly proposed using multi-temporal spectral data with given weight to estimate yield using contribution ratios of different stages to improve yield estimation, then use stepwise regression to build models for yield estimation. The contribution ratio is calculated by principal component analysis respectively. Result show that this methodology reflects the crop growth status, physiological characters at different stages, and improves the accuracy obviously.

Remote sensing of vegetation liquid water has important application in agriculture and forestry. The spectral index or features of water absorption in NIR and SWIR have been found useful for the detection of plant water content (PWC). It is unfortunately that the foliar liquid water absorption is superposed by the atmospheric vapor absorption, and it is very difficult to distinguish the contribution of foliar liquid water and atmospheric vapor on the water absorption spectral features. In this study, the novel methods using red edge width, which located outside of water absorption bands, were proposed for deriving winter wheat PWC from field canopy spectra. The significant positive correlation coefficients were observed between PWC and spectral reflectances in 740-900nm region for all the 6 growth stages. And this mechanism also affects the spectral red edge in 680- 740nm region. The correlation coefficients between PWC and red edge width derived from inverted gaussian model is significant at 0.999 confident level, and we established the statistical models for PWC by red edge width in all the 6 growth stages.

Spectral data are strongly related to canopy parameters, which are related to final yield at a critical stage of the crop growth, so it is possible to estimate wheat yield from spectral data. In this paper, firstly, the yield data were related to the field hyperspectral reflectance data in 8 different growth stages from regreening stage to milking stage. The statistical results show that yield is positive correlative with spectral reflectance in NIR bands and negative correlative in visible and SWIR bands. Secondly, the normalized difference spectral indices defined by [860 nm, 1200 nm], [890 nm, 980 nm], [820 nm, 1650 nm], [820 nm, 1650 nm] and [820 nm, 2200 nm] are designed and related to yields in all the 8 growth stages. Compared with NDVI, these spectral indices can predict yield earlier and more reliably. Finally, the remote sensing models in different growth stages for yield were built by [820 nm, 1650 nm] weak water absorption index.