The proper use of medical personal protective equipment (MPPE) is critical for frontline healthcare workers (HCWs) to handle highly contagious diseases. Due to the complexity of PPE donning and doffing protocols, public health organizations typically recommend having trained observers monitor the entire PPE donning and doffing process, preventing self-contamination and transmission. However, the high costs of manual monitoring impede the implementation of this practice, which makes AI-assisted PPE monitoring highly valuable. Some studies have applied computer vision techniques to PPE monitoring, but they have only focused on limited integrity checks at donning completion, which is unable to provide real-time warnings for abnormal actions during the doffing process. Furthermore, model practicality and user-friendliness are constrained by the lack of explainability. To address this, we propose an explainable and fine-grained dataset for MPPE doffing monitoring called the XFMP dataset. The dataset contains 3596 expert-annotated samples over three sub-tasks: doffing stage classification (DSC), abnormal action recognition (AAR), and critical region localization (CRL). Accordingly, we introduce multi-dimensional evaluation metrics for XFMP and a multitask human behavior semantic attention network (MHBSAN). Experiments demonstrate that MHBSAN outperforms alternative approaches, achieving 0.968/0.855 accuracy for stage/action classification and 0.791 Top-1 [email protected] for CRL sub-task. Moreover, it demonstrates exceptional adaptability across different healthcare environments. Ablation studies and case analyses further validate the contributions and efficacy of the proposed model regarding classification and explainability.

Wildfire spread is affected by various factors including weather, fuel, topography, and human intervention. Previous studies have focused on wildfire probability modeling for purposes of wildfire management, with less attention paid to potential wildfire behavior characteristics such as wildfire speed and intensity. Remote sensing technology has excellent advantages in deriving the characteristics and fuel variables. This study aimed to model these characteristics for wildfire hazard assessment in the Yunnan Province of China. The random forest (RF) model and the extreme gradient boosting (XGBoost) model, were selected to establish the potential wildfire behavior characteristics (PWBC) models based on explanatory variables. The results verified that elevation, fuel moisture content, and infrastructure variables played a more significant role in the models. The RF-based models performed better than the XGBoost-based ones with higher overall accuracy (≥0.83) and kappa coefficient (≥0.79), indicating the effectiveness of predicting potential wildfire behavior characteristics to assess wildfire hazards.

Accurate estimation of potential fire intensity (PFI) can improve wildfire management. The PFI can be simulated by fire spread models, but with immeasurable uncertainties. There are also some difficulties in estimating PFI with multi-source drivers, since the fire spread is limited by fire suppression. This study aimed to estimate the probability density of PFI over southwestern China, using time-series fuel and weather data as well as topographic data. The Quantile Regression and Bi-directional Long Short-Term Memory were selected to establish the prediction model of PFI. The results showed that the QR-BiLSTM performed best at the 90% confidence level. The modal PFI values extracted from the estimated probability density were closer to the observed values. This study suggests the potential of probability density estimation of PFI with artificial intelligence, for which improves wildfire risk assessment.

Dead fuel moisture content (DFMC) was usually involved and being an important part in predicting ignition potential, fireline intensity, flame length, and rate of spread. Previous studies focused on model development and paid little attention to large-scale DFMC forecasting using these models, especially process-based models. In this study, we forecast spatial 1-h (fuel with a diameter less than 0.635 cm) and 10-h (fuel with a diameter between 0.635 cm and 2.54 cm) DFMC in 16 days with meteorological variables interpolated from Global Forecast System (GFS). First, we interpolated meteorological variables including air temperature (Tair), relative humidity (RH), wind speed (Ws) and precipitation (P) to 2 km from GFS data (0.25°) for each site with DFMC measurement in Liangshan Yi Autonomous Prefecture. Then, we forecasted DFMC hourly for 96 sites with three process-based models (Simard, Nelson and fuel stick moisture model (FSMM)). Our results show that the FSMM forecasted more accurate DFMC values (1-h: R2=0.54, RMSE=6.58%, MAE=5.63%; 10-h: R2=0.73, RMSE=3.7%, MAE=2.7%) compared to the Nelson and Simard model. Our results suggest that accurate DFMC forecasts from GFS data based on our methods can be used for fire risk assessment and fire behavior prediction.

Understanding the extent to which climate affects interannual wildfire variability is key to project and mitigate wildfire. However, obtaining a robust quantification of the climate-wildfire relationship at large spatial scales remains challenging. This study employed hierarchical Bayesian framework to estimate the effect of drought on forest wildfire frequency in subtropical China. We quantified the drought-wildfire relationship across the subtropical China that the probability of excess wildfire (i.e., above normal wildfire activity level) shown disproportionate growth from 2.4% to 76.7% when vapor pressure deficit (VPD) increased from -3 to 3 (z-score). The extreme wildfires only occurred when VPD exceeds 0.5 (z-score). Our results suggest that Bayesian hierarchical model performs better in quantifying the impact of drought on wildfire by taking into account spatiotemporal heterogeneity of climate-wildfire relationship.

Wildfires pose an increasingly serious threat to human life, property, and the environment. To enhance wildfire monitoring, remote sensing satellite data has become widely used. The static meteorological satellite GK2A has emerged as a research hotspot in this field due to its advantages of all-day coverage, high resolution, and real-time data transmission. In this study, we developed a near-real-time wildfire monitoring model using GK2A remote sensing satellite data and the random forest method. The model algorithm was validated using the Xichang fire event and demonstrated high accuracy and robustness in wildfire monitoring, thereby providing important support for timely warning and response to wildfires. With the advancement of remote sensing technology, the utilization of GK2A satellite data and advanced machine learning algorithms is expected to provide more accurate and efficient methods for early detection and response to wildfires. The overall precision of the experimental results is 0.93, and the F1 score is 0.62. This study significantly contributes to improving the level of wildfire monitoring technology.

P-band PolInSAR has the potential to map forest height and biomass at the global scale with the upcoming BIOMASS mission. However, because of the strong penetration of P-band and temporal decorrelation of repeat-pass observations, volume temporal decorrelation (γVT), ground temporal decorrelation (γGT), and residual ground scattering (mmin) largely influence forest height inversion. By integrating the random volume over ground (RVoG) model and Sum of Kronecker Products (SKP) decomposition, this study proposed a multi-baseline forest height inversion method to remove the joint influence of these problems. The theoretical simulation and empirical experiments using airborne P-band PolInSAR data in tropical forests, Nouragues, explicitly explored how each of the three problems affects the inversion. γVTandmmingenerally affect forest height inversion more than γGT, and the concurrence of them brought severe overestimation to RVoG inversions (RMSE ranges from 7.6 to 18.8 m). Stepwise comparisons show that compensating for each of γGT, γVT, andmmincould improve the inversion accuracy further. The proposed multi-baseline inversion simultaneously solved all three problems and produced the best accuracy (RMSE of 3.4 m), and it has a stable performance for another two more different multi-baseline datasets, producing similar inversion accuracies (RMSE of 3.3 and 3.6 m). The additional experiment using BIOSAR 2007 datasets with varied temporal baselines of 0 day, 30 days, and 56 days demonstrated that the proposed method has stability against temporal decorrelation. Consequently, the proposed method improved both the accuracy and robustness of forest height inversion.

Forest tree species are an important factor in stand structure surveys, which are the basis for forest ecological planning and the development of related forest policies. Since the number and types of tree species are related to ecosystem parameters such as biodiversity and habitat quality. An accurate and detailed assessment of tree species is a prerequisite for effective ecosystem management. Complex mountainous areas are surrounded by clouds all year round, making it difficult to obtain clean and usable images. Change Continue Detection and Classification(CCDC) is a method for land cover change detection and is able to produce synthetic Landsat images at any moment based on the time series model. The purpose of this paper is to verify the feasibility of synthetic Landsat imagery for tree species classification. In this study, a harmonic function model was built by using 518 images of northern Sichuan province of China from 2000 to 2020. The harmonic function model was used to generate synthetic Landsat images, then, three ensemble classification models were used to classify the trees based on the synthetic images. The total accuracies of the three classifications were 85% (RF), 60% (AdaBoost), and 78% (GBDT), respectively. The results show that Landsat synthetic images can be well applied to tree species classification in complex environments.

Forest biomass is a complex parameter that is a function of multiple forest structure parameters, such as forest height, DBH, and woody density. At present, forest height is most likely to be estimated over the globe from spaceborne remote sensing, particularly with the upcoming BIOMASS mission. This study inverted forest height using P-band repeat-pass multi-baseline PolInSAR data and evaluated the robustness across three different forest sites. Afterward, the PolInSAR height was converted into forest AGB. Results show that the method produced a more accurate forest height in tropic forests than that in boreal forests, but there is little bias for forest height estimates across three sites, with$\mathrm{R}^{2}$of 0.86 and RMSE of 3.1m in contrast to LiDAR forest height. One power equation was found able to convert PolInSAR forest height into forest AGB across three sites, producing$\mathrm{R}^{2}$of 0.85 and RMSE of 51.8 tons/ha using LOO cross-validation, which is at a similar level to the AGB accuracy estimated using LiDAR-derived forest height$(\mathrm{R}^{2}$of 0.87 and RMSE of 47.9 tons/ha).

Temporal decorrelation is a critical issue to be addressed for forest height inversion using repeat-pass Polarimetric SAR Interferometry (PolInSAR) data. Temporal decorrelation plus the residual ground contribution will bring a 2-D ambiguity to volume-only coherence, which makes the inversion underdetermined. This study utilizes the overlapped area of the ambiguities from different baselines to reduce the ambiguous region and to more accurately retrieve the volume-only coherence and forest height. The method was validated and compared with two commonly used single-baseline inversions, moreover, its robustness was evaluated across two tropic forest sites, Paracou and Nouragues, French Guiana. Results show that the presented multi-baseline inversion scheme successfully reduced both the influence of temporal decorrelation and ground contribution, and performed better than any of the single baseline used. For Paracou, R2 was improved from 0.35 to 0.42, RMSE decreased from 4.7 to 2.7m, and ubRMSE decreased from 3.0 to 2.7m; for Nouragues, R2 was improved from 0.43 to 0.56, RMSE decreased from 5.8 to 5.7m, and ubRMSE decreased from 5.6 to 5.0m, at a fine validation scale of ca. 20×30m.

With global warming, the number and the burning area of wildfires will continue to increase. Wildfire risk assessment is to assess the occurrence probability and spread risk of wildfire by combining fire risk factors, like weather, fuel and terrain. In the past, linear regression models and general artificial intelligence models were used to assess wildfire risk. But the wildfire datasets are typical non-equilibrium, using these methods often results in a bias towards non-fire class, leading to lower point in fire class recognition. Algorithms for imbalanced datasets, such as balanced random forest, can improve the identification accuracy of the minority class, but are rarely used in wildfire risk assessment. In this study, we discussed the performance of random forest, extreme gradient boosting and balanced random forest models in wildfire datasets of the Great Xing'an Mountains. The results showed that: balanced random forest model greatly improves recall and f-score, and its wildfire risk distribution map shows that high fire risk areas are more accurate representation of the location of wildfire. Therefore, this study believes that balanced random forest model is more suitable for wildfire risk assessment.

Forest fuel load plays an important role in fire ignition, spread and intensity. Accurate spatial distribution information of fuel load is vital for fire managers to make decisions. However, most of the existing studies using spectral or texture information of optical data, which is seriously affected by atmospheric conditions with limited penetration. SAR with all-day and all-weather work characteristics provides a favorable opportunity to estimate fuel loads in Southwest China with cloudy and foggy. Moreover, SAR can penetrate through leaves to branches and stems, which reflects the forest vertical structure well. However, there was little research applying SAR data to fuel load estimation. In this study, we focus on forest above ground live fuel load estimation (biomass fractions), including stem fuel load (SFL), branch fuel load (BFL) and foliage fuel load (FFL). We explored the potential of dual polarimetric data, Sentinel-1 for coniferous forest fuel loads estimation in the Southwest of Sichuan, China. To understand scattering mechanisms at C-band in Pinus yunnanensis forest, the Michigan Microwave Canopy Scattering (MIMICS) radiative transfer model was used. And Multiple Linear Regression (MLR) method was used to estimate fuel loads. Results show that VH and VV polarizations were both sensitive to three types of fuel load. Combined with the simulation of MIMICS, we found VH polarization was more sensitive to FFL while VV was more sensitive to SFL. Additionally, Sentinel-1 performs well in all three types of fuel load estimation (FFL: R2= 0.52, RMSE=1.43 Tons/ha; BFL: R2=0.58, RMSE=1.88 Tons/ha; SFL: R2=0.57, RMSE=2.97 Tons/ha), indicating that Sentinel-1 data has great potential in FFL estimation and fire prevention.

Dead fuel moisture content (DFMC) is of important significance for estimating and predicting forest wildfire risk, in which 1-h dead fuel is most critical as the easiest fuel to ignite. Current methodologies based on empirical and physical models rely heavily on meteorological data from uneven and sparse stations. In contrast, satellite data become a better choice since its continuous surface observations. Of all satellites, Himawari-8 is the most appropriate data to meet the rapid changes of DFMC throughout the day owing to its high time resolution. Thus, this study explored the application of 1-h DFMC predicting using machine learning from Himawari -8 data. Random forest was selected for the prediction and linear regression was used for comparison. The results showed that random forest has a satisfactory performance with higher R2 (0.53) and lower RMSE (3.15%) than that of linear regression (R2=0.21, RMSE=5.47%). The research suggested that predicting 1-h DFMC at regional scales using machine learning from Himawari -8 data is promising.

Aboveground biomass (AGB) is an important indicator of grassland state. Remote sensing estimation method of grassland biomass can provide important decision support for decision-makers and relevant personnel on grassland management, such as the rational development and utilization of grassland resources, the local ecological environment protection and the rational development of animal husbandry. In this study, based on Landsat 8 and Sentinel-2 A/B satellites data, PROSAILH radiative transfer model (RTM) and look up table (LUT) algorithm were applied to temporally retrieve and map the grassland AGB in Qinghai Province from September 2019 to August 2020, an important graziery region in China. Further analysis on the relation between the AGB and meteorological data showes that the precipitation and air temperature in Qinghai is consistent with the average biomass dynamic, indicating the considerable effect of the meteorological factors on the AGB in this region.

Forest Dead Fuel Load (FDFL) is vital for fire prevention and suppression since it affects the surface fire ignition and intensity directly. The accurate spatial distribution information of FDFL can provide decision support for fire managers. Remote sensing (RS) technology is a unique way to estimate FDFL on a large scale. However, most researches focus on the application of LiDAR data which is expensive to analyze fuel dynamics on spatiotemporal scale. Little attention has been devoted to freely accessible datasets such as Sentinel-1 and Sentinel-2 with global coverage. This study not only combined these two data but also the site conditions (i.e., elevation, slope and aspect) to estimate the FDFL in the southwest of Sichuan, China. The machine learning method, Random Forest Regression (RFR) was selected to manage the multiple and nonlinear relationships between RS data and FDFL. Results show that 1h and all FDFL (the sum of 1h, 10h, 100h and litter) can be more indicated by RS data (1h:$\mathrm{R}^{2}=0.57,\ \text{RMSE}=0.18$Tons/ha; All:$\mathrm{R}^{2}=0.59,\ \text{RMSE}=1.81$Tons/ha). The representation ability of RS data for 10h, 100h and litter is relatively weaker (10h:$\mathrm{R}^{2}=0.41,\ \text{RMSE}=0.47$Tons/ha; 100h:$\mathrm{R}^{2}=0.40,\ \ \text{RMSE}=1.07$Tons/ha; litter:$\mathrm{R}^{2}=0.29,\quad \text{RMSE}=1.66$Tons/ha). Hence, this study demonstrated the potential of multi-source RS data for FDFL estimation.

P-band PolInSAR is potential to map forest height and biomass at the global scale with the upcoming BIOMASS mission. However, there are three major problems in forest height inversion using P-band repeat-pass PolInSAR data, i.e. volume temporal decorrelation, residual ground scattering, and ground temporal decorrelation. Based on the random volume over ground (RVoG) model and the Sum of Kronecker Products (SKP) decomposition, this study developed a multi-baseline forest height inversion method to remove the joint influence of these three problems. Based on the TropiSAR 2009 datasets in Nouragues, results show that the concurrence of the three problems brought severe overestimation to three-stage inversion (RMSE of 7.6-18.8 m for different baselines). Stepwise comparisons show that each compensation of the three problems improved the inversion accuracy. The proposed multi-baseline inversion simultaneously removed all three problems and produced the best accuracy for both short (20 m) forest (RMSE 3.4 m), which is much better than the inversion accuracy of the used single baselines.

Most of the current optical-based forest AGB estimation methods only use spatial information of reflectance, making estimates susceptible to saturation and noise. We introduced dynamic time-series information on forest reflectance into AGB estimation and explored its ability to improve estimation accuracy. The experiments were conducted based on 10304 in situ AGB plots and corresponding Landsat 5/7/8 data from 1984 to 2019 in Australia, using a random forest algorithm for AGB estimation. Results show that time-series modeling of reflectance could remove most of the noise in observed reflectance, and improved the correlation between reflectance and AGB. After introducing the time-series parameters into AGB estimation, estimation accuracy was significantly improved, with R2 increased by 0.08 and RMSE reduced by 7.5 tons/ha.

PolInSAR technique has been proved effective in mapping forest height. However, for P band repeat-pass PolInSAR data, the influence of residual ground scattering contribution and temporal decorrelation are non-negligible for forest height inversion. To resolve these two problems, we developed a dual-baseline inversion method to reduce the influence of ground scattering contribution, and then cooperated it with Random-Motion-over-Ground (RMoG) model to compensate for the influence of temporal decorrelation. The Three-stage inversion showed a severe overestimation for the whole image, producing R2of ca. 0.4 and RMSE of ca. 7m. For near range (with$k_{z}$larger than 0.1rad/m), the overestimation was mainly caused by ground scattering contribution, which was effectively solved by using fixed extinction method. For far range (with$k_{z}$smaller than 0.06rad/m), the influence of ground scattering contribution seemed decreased because the fixed extinction method did not reduce the overestimation, and overestimation was mainly caused by volume temporal decorrelation. Besides, the influence of ground scattering contribution on forest height inversion seemed greater than that of temporal decorrelation. The proposed method successfully solved both of the two influences, and reduced the overestimation for both near and far range, producing the best accuracy with R2 of 0.62 and RMSE of 3.73m.

Wildfire is a typical complex nonlinear process, which has various interactions with weather, fuel, topography, and human variables. Previous studies on wildfire danger modeling mainly focused on the current climate or future scenarios of these variables. However, the outbreaks of wildfire are not only related to various factors at the current moment but also related to the change and trend of factors before the outbreak time point. The short-term features of weather and fuel dynamics$\text{(weather}\&\text{fuel}_{\text{short-term}_{-}\text{features}})$were therefore assumed to be of great importance in wildfire danger modeling within this context. To verify this hypothesis, the wildfire danger was modeled across the fire-prone area over southwest China from 2001 to 2020. The state-of-the-art eXtreme Gradient Boosting (XGBoost) model was applied to explore the relations between this topography, weather, fuel, and human variables and the wildfire occurrence, as well as to identify the role of five categories of$\text{weather}\&\text{fuel}_{\text{short}-} \text{term}_{-}\text{features}$, i.e., the$t^{C3}, t^{Entropy}$, Ricker's wavelet transformation$(f^{cwt})$, regular statistic description$(t^{RSD})$, and fast Fourier transform$(t^{FFT})$. Results showed that wildfire danger modeling can be improved by consideration of the$\text{weather}\&\text{fuel}_{\text{short-term}_{-}\text{features}}$which is proved to play an important role in driving the wildfire occurrence by factor importance analysis.

Wildfire, one of the most serious disasters in the world, making a huge threat to local economic development and residents' life and property safety. Therefore, it is crucial to realize near real time wildfire detection in specific high wildfire risk area. Himawari-8 is a state of art Geostationary Orbit Satellites (GOS), which can realize near-real time wildfire detection and dynamic monitoring of wildfire changes. In this study, three features strategies: spectral, spectral calculated and spectral with spatial were used to extract features from Himawari-8 and Topography data. Next, Random Forest models were trained by these features. And then, we evaluated models on Himawari-8 data generated in March 29, 2020 6:10 UTC, March 30, 2020 5:50 UTC, March 31, 2020 5:30 UTC and April 1,2020 5:50 UTC to detect specific wildfire events in Southwestern China. The results showed an overall precision of 64.24%, 64.08% and 68.07%, and an overall F1-score of 0.737, 0.677, 0.724 based on three different features strategies. This study proved that our models have the ability to detect wildfire point accurately, and model with strategy.3 performed the best when considering precision, omission and F1-score.

Chlorophyll is an essential pigment for photosynthesis of crops, which indicates the growth status of crops. Accurate and robust information on the spatial dynamics of chlorophyll content is of critical importance for crop growth status assessment and corresponding response activities. However, previous studies mainly focus on the methodologies based on in situ hyperspectral data, which are not applicative for regional chlorophyll content mapping. In this content, Unmanned Aerial Vehicle (UAV) based hyperspectral imagery, with high spatial and spectral resolution, may provide the spatial distribution of chlorophyll content accurately over crop fields. In this study, an empirical model between wavelet features, derived from UAV based hyperspectral imagery by continuous wavelet transform (CWT), and chlorophyll content (measured by a portable soil-plant analysis development meter) is proposed by support vector regression (SVR). The results suggest that the UAV based hyperspectral imagery combined with CWT demonstrates good performance in rice chlorophyll content estimation with R and RMSE of 0.81 and 3.51, respectively. Moreover, the wavelet coefficients corresponding to two bands at red (640nm, 628nm) and one band at green (548nm) are the most effective wavelet features to estimate chlorophyll content of rice.

Spatiotemporal mapping of shrublands live fuel moisture content (LFMC) is of critical significance for wildfire risk assessment. Previous studies generally estimated the shrubland LFMC based on PROSAILH radiative transfer model (RTM), the model assuming that shrub canopies were homogenously and continuously distributed. We argued that this treatment was inaccurate since shrublands present diverse distribution and structure that normally can be classified into closed and open shrublands according to IGBP scheme. To this end, this study explored the effects of PROSAILH for open and closed shrublands LFMC retrieval based on National Fuel Moisture Database. The results showed that 1) PROSAILH performed poorly on open shrublands LFMC retrieval due to heterogeneity, whereas 2) better results were found for closed shrublands LFMC cases. Thus, this research suggested that PROSAILH is not suitable for all types of shrublands LFMC retrievals.

Wildfire is one of the major environmental issue in economic, social and environmental damages globally. Three major forces are essential for understanding the wildfire occurrence, fuel, weather and topography, as regulated by the `fire environment triangle'. Within this concept, a big database that involved all the satellite-derived fuel parameters (fuel moisture content and its derived variables, Leaf area index and vegetation type), weather parameters (relative humidity, air temperature, precipitation, wind speed), satellite-derived topography parameters (slope, aspect, elevation) and corresponding historical wildfires events that calculated from the MODIS burned area product (MCD64A1) was built from 2001-2018 over China. To explore the relations between wildfires and these parameters, the random forest model was developed to mine the relations between the explanatory variables (fuel, weather and topography parameters) and the dependent variable (historical wildfires events). Results showed that the models obtained a high accuracy level in fire spatial pattern distribution, indicating reasonable capacity in predicting wildfire occurrence in China.

In this study, we focus on the controllability of fractional-order damped systems in linear and nonlinear cases with multiple time-varying delays in control. For the linear system based on the Mittag-Leffler matrix function, we define a controllability Gramian matrix, which is useful in judging whether the system is controllable or not. Furthermore, in two special cases, we present serval equivalent controllable conditions which are easy to verify. For the nonlinear system, under the controllability of its corresponding linear system, we obtain a sufficient condition on the nonlinear term to ensure that the system is controllable. Finally, two examples are given to illustrate the theory.

Detecting objects becomes an increasingly important task in very high resolution (VHR) remote sensing imagery analysis. With the development of GPU-computing capability, a growing number of deep convolutional neural networks (CNNs) have been designed to address the object detection challenge. However, compared with CPU, GPU is much more costly. Therefore, GPU-based methods are less attractive in practical applications. In this article, we propose a CPU-based method that is based on convolutional support vector machines (CSVMs) to address the object detection challenge in VHR images. Experiments are conducted on three VHR and two unmanned aerial vehicle (UAV) data sets with very limited training data. Results show that the proposed CSVM achieves competitive performance compared to U-Net which is an efficient CNN-based model designed for small training data sets.

Paddy rice is one of the most important crops in the world. Information of the area and spatial distribution of paddy rice is significant for food security, water resources management, and greenhouse gas (methane) emissions. Paddy rice field is characterized by an initial period of flooding and transplanting, during which period open canopy (a mixture of rice crops and surface water) exists. The present algorithms for paddy rice recognition are mainly based on the unique physical features of paddy rice and the vegetation indices, which are sensitive to dynamics of the canopy and surface water content. In this study, the rice growth calendar data and the cropping pattern (e.g., single rice or double rice) information were used to improve present phenology-based rice recognition approach by determining the potential temporal window of flooding and transplanting periods over a year. Other land cover types (e.g., snow, evergreen vegetation and permanent water bodies) with potential influences on paddy rice identification were removed (masked out) due to different temporal profiles. The results showed that the estimated rice area is in line with the provincial agricultural statistics with R2and slope are 0.951, 1.049, respectively, and the relative errors of the different province are between -15.81% to 25.75%.

Semantic segmentation is understanding images at pixel level, which becomes increasingly vital in unmanned aerial vehicles (UAV) imagery classification tasks. With the powerful calculating ability of GPU, a growing number of deep convolutional neural networks (DCNNs) are designed to address semantic segmentation challenges. However, compared with CPU, GPU is much more costly, and GPU relies on powerful supplementary equipments to support it. Therefore, GPU-based methods are hard to carry out in practical applications. In this study, we propose a CPU-based method named fully convolutional support vector machine (FCSVM) to address the semantic segmentation challenge in UAV images. On the one hand, we adopt the SVM kernel from the original convolutional SVM networks (CSVM), which completes classification at image level. On the other hand, FCSVM consists of two processes which are a compressed process and a extensive process. In the compressed process, the FCSVM has convolutional layers and reduction layers. In the extensive process, the FCSVM has convolutional layers and upsampling layers. This structure allows FCSVM to classify images at pixel level using limited number of training data. The experiments are implemented on one UAV dataset with very little training data. The result shows our FCSVM achieves competitive performance compared to modern state-of-the-art semantic segmentation methods.

Fuel Moisture Content (FMC) is a key indicator for assessing wildfire risk and fire spread rate since fire normally broke out under low FMC condition. However, how the FMC driving wildfire activities is still unclear. To understand the relation between FMC and fire occurrence, this study focused on analyzing the FMC retrieved from Radiative Transfer Model (RTM) and historical fires extracted from MODIS Burned Area (BA) product MCD64A1 to determine critical FMC thresholds for wildfire occurrences in southwest China. The results indicated that there were three significant FMC thresholds for forest: 138.9%, 119.0% and 53.1%, and three thresholds for grassland: 138.3%, 70.2%, and 10.6% that associated with the occurrence of wildfires. Large fires and high frequency of fires became vulnerable when FMC fell below those thresholds.

Fuel moisture content (FMC) is a critical variable in assessing wildfire risk and its behavior. Previous studies normally focused on the methodologies based on optical remote sensing data for FMC retrieval. However, active microwave technique, which processes the advantage of high sensitivity to surface moisture, all-weather and all-time work capability and strong penetrability, attracted more attention in surface parameter monitoring, particularly for the polarimetric SAR which provides more sufficient object scatter characteristic. In this paper, we retrieved the FMC for a grassland based on the multiple linear regression analysis of polarimetric decomposition parameters from Radarsat-2 data. The results show that the correlation coefficient (R) and root mean square error (RMSE) reached to 0.658 and 30.319% when compared to the measured FMC. Finally, the presented method was used for spatial and temporal mapping of FMC in the target study area.

Spatiotemporal monitoring of fuel moisture content (FMC) is vital to assessing the wildfire risk and its behavior. Optical remote sensing data-based FMC estimation have been wildly explored in previous studies. However, limited studies focused on FMC retrieval from the active microwave technique represented by synthetic aperture radar (SAR) data, which processes the advantage of higher sensitivity to surface moisture and better all-weather and all-time work capability than optical data. This is the first study to assess the performance of time series dual-polarization Sentinel-1A data for FMC estimation from coupled the bare soil backscatter Linear Model and the vegetation backscatter Water Cloud Model. The results show that the simulated backscattering coefficients and FMC are in line with the measured Sentinel-1A data and FMC with R2and RMSE are 0.549, 0.354 dB, and 0.543, 13.579 %, respectively.

Spatiotemporal simulation of Fractional Vegetation Coverage (FVC) is of great significance for the protection and management of the ecological environment. In this study, the growing season FVCs of the South Qilian Mountains from 2000 to 2017 were extracted from the MODIS vegetation indices product (MOD13Q1), and then that were used to train the Back Propagation (BP) artificial neural network to estimate the annual FVCs of the South Qilian Mountains in the next 7 years (2019-2025). The results show that the established model has a good performance through verification FVC data in 2018, with the coefficient of determination (R2) is 0.9462 and the root mean square error (RMSE) is 0.0118. The simulation results of the model indicate the FVC will present a trend of growth in the following years. This study indicates that the combination of BP neural network and remote sensing data can effectively simulate the spatiotemporal pattern of fractional vegetation coverage, which can further contribute to the environmental protection.

In this study, the burn severity of several wildfires ignited at northern Australian tropical savannas area were estimated using the Forest Reflectance and Transmittance (FRT) radiative transfer model (RTM) and Sentinel-2A Multi-Spectral Instrument (MSI) satellite data. To alleviate the spectral confusion between severe (SV) and not-severe (NSV) burnt levels caused by sparse tree distribution, the MODIS Vegetation Continuous Fields (VCF) tree cover percentage data was used to constrain the inversion. The results showed that the accuracy of burn severity estimation significantly improves when considering the tree coverage, with overall accuracy for two study sites increasing from 65% to 81% and kappa coefficient from 0.35 to 0.55. Future work will focus on extending the methodology to other ecosystems.

Tianjin city, China has seen rapid urban expansion and economy development especially since the Chinese reform and opening up in 1978. Understanding their internal relationship is important for urban management. Considering of the long-history records with medium spatial resolution (30 meters), Landsat time series (LTS) has become a key remote sensing dataset for monitoring urban changes. As the impervious surface is an important indicator for assessing urban environment, in this study, we extracted the impervious surface maps in Tianjin between 1990 and 2017 based on LTS using a Continue Change Detection and Classification (CCDC) algorithm. On the other hand, we quantitatively explored the relationships between the impervious surface and Gross National Product (GDP). Results show that the impervious surface expansion in Tianjin experienced two stages between 1990 and 2017 with the area increased by 521.95 km2, and the correlation coefficient between the area of impervious surface and GDP value was as high as 0.9718.

Nonvolume decorrelation (γNonvol) united with the unknown ground contribution will bring a 2-D ambiguity to volume-only coherence, making the inversion underdetermined even when multiple baselines are available. In the context of random volume over ground (RVoG) model and three-stage algorithm, this paper theoretically presented the varied response of different baselines to both γNonvoland ground contribution, and then proposed a new multi-baseline inversion method to reduce the 2-D ambiguity. The proposed method includes two steps, calculating the common overlapped ambiguity from different baselines and fixing the extinction coefficient, to more accurately retrieve the volume-only coherence and forest height. It makes no assumptions on γNonvoland ground contribution. The method was validated and compared with three single-baseline inversions and two published multi-baseline inversions by using the airborne P-band polarimetric SAR interferometry (PolInSAR) data and the reference data of LiDAR canopy height model (CHM) over a dense rainforest site. Results showed that the developed multi-baseline method successfully reduced the combined influence of both γNonvoland ground contribution, and performed better than any single baseline, improving the R2from 0.60 to 0.77 and unbiased root-mean-square error (RMSE) from 1.32 to 1.04 m at the scale of ca. 100 × 140 m2. Moreover, the multi-baseline scheme is relatively robust among different baseline combinations.

The change of water resources under global warming is one of the most important issues faced by national economy and social development. Accurate estimation and analysis of Terrestrial Water Storage Anomaly (TWSA) are important for understanding global and regional water cycle process. Gravity Recovery and Climate Experiment (GRACE) satellite data is used to inverse water storage change in China and the correlation between it and precipitation from Tropical Rainfall Measuring Mission (TRMM) is also analyzed. The study shows that GRACE TWSA and TRMM precipitation are consistent and linear. In the best regression model, R2and RMSE are 0.8025 and 11.1469 respectively, which suggests that precipitation is the source of water supply and also the main impact of water storage change.

Dam construction is one of the solutions to the problem in water resources distribution and utilization caused by global climate change and the calculation of water energy reserves can provide scientific decision for it. The emergence of remotely sensed data and the application of Geographic Information System (GIS) make water reserves calculation a reality on a large scale. Hydro30 and measured runoff data are adopted to extract indicators of water energy reserves and the analysis of its spatial distribution is carried out in Anabar basin, Russia. The result shows that the most suitable site for waterpower development is concentrated in the upper reaches and the reserves of main rivers are larger than that in the source and tributary rivers.

Images classification is an essential field in remote sensing community. However, a variety of target shapes, as well as changing conditions during multiple time periods and different areas usually result in shifts in classification. This problem can affect classification results seriously. Although the affection is significant in remote sensing classification, very few people have considered this issue, and have solved it. In this paper, we introduce the associative domain adaptation (ADA) method to address this challenge. We apply this algorithm to two public remote sensing datasets. One is famous UC Merced dataset; another is NWPU-RESISC45 dataset which has a much more variance within the class. We then build a classification model by using UC Merced training images and labels as well as using training images from NWPU-RESISC45. This semi-supervised classification performance achieves an impressive test accuracy on the NWPU-RESISC45 test dataset.

Fire Spread Rate (FSR) is one of the key factors for fire rescue and prevention. Remote sensing images have an advantage of acquiring intuitive information timely. To achieve extracting real-time FSR from remote sensing data, a method based on the movement rate of burned area centroid is presented in this study. The FSR extracted from geostationary Himawari-8 (H-8) data in two bushfires outbroke in Esperance, Western Australia in 2015. And the FSR estimated from CSIRO (Commonwealth Scientific and Industrial Research Organization) Grassland Fire Spread Model (CGFSM) were set as the benchmark to assess the presented approach. The results illustrated that the proposed method yield a promising accuracy by comparing with the FSR from CGFSM in that two fires, with the coefficient of determination (R2) reaches to 0.76 and root-mean-square error (RMSE) is 0.50 m·s-1. Furthermore, this study provides a potential application of geostationary satellite in extracting real-time wildfire behavior.

Fuel moisture content (FMC) is a critical factor in assessing wildfire risk and its behaviour. Traditional field measurement of this variable is time-consuming and is impossible to extend to large-scale and dynamic applications. The canopy water has strong absorption characteristic in near and shortwave infrared spectra, allowing the near-real-time, multi-temporal and -spatial estimation of the FMC from remotely sensed data available. During last decade, numerous statistic- or physical model-based studies were carried out for the estimation of this variable. As FMC is responsive to weather variations, diurnal determination of this variable is essential for wildfire early-warning. With the launch of Himawari-8 in 2014, 10 mins images are available from this satellite, making real-time retrieval of the FMC achievable. Thus, this is the first study to retrieve diurnal FMC from Himawari-8 images, with the purpose for real-time wildfire risk assessment in near future.

This study focuses on the forest fire risk assessing using entirely remote sensing derived variables. These variables include Fuel moisture content (FMC), Normalized Difference Vegetation Index (NDVI), Leaf Area Index (LAI), Elevation and Slope. The Difference and Anomaly of FMC in time series are also calculated since FMC is one of the critical factors in assessing the wildfire risk. The logistic regression model is used to integrate all the variables in the fire occurred and none-occurred areas to derive the Fire Risk Index (FRI). A case study of the above methodology is applied to assess the FRI in Yunnan Province in China. The result shows that the AUC is to 0.8 for grassland and 0.81 for woodland, which indicates the good performance of the presented methodology in assessing forest fire risk.

Desertification is a severe issue that has received increasing attention currently. This paper aims at the desertification degree assessment and trend analysis between 2000 and 2014. Firstly, based on the negative correlation between albedo and Normalized Difference Vegetation Index (NDVI), desertification difference index (DDI) products are completed, which are applied to assess desertification degree. For the purpose of monitoring the tendency of global desertification, the time-series data from MODIS are applied to work out the normalized yearly albedo and NDVI. Then, based on DDI products and the trend analysis approach, the global desertification tendency products are produced. Preliminary analysis showed that the ranking of global desertification seriousness relates to the temperature zone. Desertification trended to be aggravated in central North America, eastern Asia, the Pacific Rim, and central Australia. The results above may promote the scientific decision, and help prevent and control the desertification.

With the development of machine learning, many researchers have used machine learning models for building extraction in high resolution remote sensing images. Especially for the recently proposed deep learning models, it has been widely used for building detection in the urban monitoring. In this study, the performances of images segmentation for building extraction based on Fully Convolutional Networks (FCN) model and shallow models are qualitatively and quantitatively compared. Firstly, the public aerial dataset of Massachusetts building dataset[1] are preprocessed to extract features. Then, we trained shallow models and deep model on Massachusetts building dataset. Moreover, the trained FCN model and shallow models are used to extract the building on the same image. Finally, we compared performances between shallow models and deep model. It is found that FCN gives the highest recall 0.63, precision 0.62, and F-measure rate 0.63. The qualitative and quantitative analysis of the building extraction results fully demonstrates that FCN gives the best performance compared with traditional shallow models.

In this study, an alpine watershed was selected, and four kinds of biophysical parameters (Leaf Area Index, LAI; Canopy Water Content, CWC; Canopy Chlorophyll Content, CCC; Fractional Vegetation Cover, FVC) were retrieved and their products were generated at the regional scale from 2013 to 2014. PROSAIL radiative transfer model and modified mixed sub pixel model were used to estimate these biophysical parameters based on Landsat 8 OLI surface reflectance data. The results suggested that the estimated biophysical parameters yielded a promising accuracy compared with the ground measurements. The inter-partition statistic was also conducted by analyzing the grassland peak growing season (July and August), and confirmed that the pasture yield and water stress condition of grassland in Lake Qinghai watershed of 2014 is better than 2013.

Solar energy (SE) is accepted as a key resource for easing the tense situation of global energy supply. It is urgent to figure out the potential for global solar energy utilization. In this paper, a multicriteria evaluation (MCE) model was constructed by comprehensively integrating resources richness, resources stability, resources utilization value, and natural environment together with weights. The resources richness, resources stability and resources utilization value derive from ERA-Interim reanalysis datasets with spatial resolution 0.125°. The natural environment comes from MODIS product MCD12C1 with spatial resolution 0.05°, and global multi-resolution terrain elevation data 2010 (GMTED 2010) with spatial resolution 30". Besides, the global solar energy utilization potential map was produced, which shows most of the high potential region located in north of African. Additionally, 3% and 75% large photovoltaic (PV) power plants (LPVPP) falls on high and medium-high potential regions respectively.

The retrieval of soil moisture in vegetated areas is often hampered by the influence of vegetation and surface roughness. The polarimetric SAR (PolSAR) can provide complete information related with the electromagnetic scattering characteristics of natural targets, which may contribute to the soil moisture in vegetation covered areas. The objective of this study is to evaluate the potential of Radarsat-2 polarimetric parameters for helping soil moisture retrieval in prairie areas. The ground measurements and satellite data collected from the Ruoergai prairie are used to test and validate the experiments. From the results of soil moisture retrieval, it is validated that the Radarsat-2 polarimetric parameters are sensitive to the variation of the soil moisture. Therefore, the polarimetric parameters are very helpful for estimating the soil moisture in prairie areas.

The Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER) imagery provides a more subtle data to find mineralized alteration zones. Through the analysis of spectral characteristics of minerals that associate with hydrothermal alteration with the reference spectral from the USGS library, we combine the band ratio method and the Principal Component Analysis (PCA) to extract iron oxides and hydroxyl. The result shows that band4/band1 and band2/band1 highlights the iron oxides, band4/band6 and band5/band7 show Al-OH and Mg-OH of hydroxyl respectively. A case study of the presented method was applied to mineral alteration extraction in northeast of Sichuan province, China. Furthermore, Spectral feature fitting (SFF) was applied to test the mineral alteration points we obtained. It demonstrates that the combination of band ratio and PCA method is effective in extracting remote sensing minerals alteration.

Landslide susceptibility assessment is necessary for disaster management and safe planning in mountainous regions. In this paper, we have used rough set theory to select ten landslide-related factors (slope, aspect, elevation and so on) derived from maps of geology and remote sensing data, and produced landslide susceptibility map in Beichuan Country, Sichuan Province of China based on a model integrating C4.5 decision tree and m-branch smoothing. The performance of method was evaluated by receiver operation characteristic (ROC) curve and the area under the curve (AUC). The result shows that the above initial controlling factors are very important for landslide activity in study area. When the parameter value of m-branch smoothing was set as 14, the method can get the best prediction accuracy with AUC value of 87.16%. These results demonstrated the proposed method for landslide susceptibility assessment is quite effective and can be applied to more disaster assessment problems.

This paper examines the potential of Radarsat-2 C-band synthetic aperture radar (SAR) data for quantifying the spatial variability of soil moisture during the agriculture growth period. To remove the effect of crop within total backscattering, a method that adequately represents the scattering behavior of vegetation-covered area by defining the scattering of the vegetation and underlying soil was developed. The Dubois model was employed to determine the backscattering from the underlying soil. The modified Water Cloud Model was used to reduce the effect of backscattering caused by the vegetation. Soil moisture was derived by the inversion scheme which uses of the dual polarizations (HH and VV) available from the quad polarization Radarsat-2 data.

The robust and accurate retrieval of biophysical vegetation variables from remotely sensed data using radiative transfer models (RTMs) has been generally hampered by the ill-posed inverse problem. To alleviate this issue and improve the retrieval accuracy, a two-step inversion method based on the PROSAIL (PROSPECT+SAIL) radiative transfer model has been developed to consider the spectral information from neighboring pixels (3 × 3 windows) and the spatial dependency between these pixels. A case study was conducted to retrieve canopy water content (CWC) values from Landsat-8 OLI data for a plateau grassland located in the northeastern Qinghai province, China. The result showed that the proposed inversion method that the combination of nearby information yields more accurate results than the standard pixel-based inversion methods. With the proposed method, the coefficient of determination R2reached 0.82 and root-mean-square error (RMSE) was 65.89 g·m-2, which was superior to the standard pixel-based inversion method with the R2= 0.77, and RMSE 76.58 g·m-2.

Leaf area index (LAI) is an important biophysical vegetation variable in many models describing vegetation atmosphere interactions. This study presents a method to assimilation leaf area index(LAI) obtained by a simple downscaling technique into the World Food Studies(WOFOST) to improve simulation LAI in time series in plateau grasslands located in northeast of Qinghai province, China. Using the Sobol' global sensitivity analysis approach, serval local and regional crop parameters of WOFOST model were identified to be recalibrated. Then, the ensemble Kalman filter algorithm was introduced to assimilate the 30m LAI into the WOFOST model to update the model state for improving simulation effect.

The distribution of soil moisture is important for modeling hydrological and climatological processes to understand the Earth energy cycle and balance. The major difficulty for soil moisture retrieval in vegetated areas is how to separate the individual scattering contribution of soil moisture, vegetation, and surface roughness from the backscattered radar signal. In this paper, a semi-empirical method was proposed to retrieve soil moisture in the Ruoergai prairie using single temporal Radarsat-2 data. It was formulated by integrating the advanced integral equation model (AIEM), a semi-empirical ratio vegetation model, and optimum surface roughness parameters. The AIEM was run in the forward mode to simulate the backscattering coefficient of bare soil surface. The ratio vegetation model was applied for eliminating the vegetation effect from the observed backscattering coefficient. Meanwhile, four different vegetation parameters were used to characterize the change of vegetation: leaf area index, vegetation water content, normalized difference vegetation index, and enhanced vegetation index. A global search method was used to find out the optimum surface roughness parameters, which made the relationship between the observed and retrieved soil moisture reach up to the best. The collected in situ measurements and satellite data from the Ruoergai prairie were employed to validate the feasibility and effectiveness of the introduced method. From the analysis of experiment results, the optimum surface roughness parameters had a relative change when different vegetation parameters were used to parameterize the ratio vegetation model. In addition, the root-mean-square height had a significant impact on the accuracy of soil moisture retrieval compared with correlation length. The best retrieval result was obtained when EVI was used to remove the influence of vegetation, with a correlation coefficient of 0.84 and root-mean-square error of 4.05 vol · %. Therefore, compared with other three vegetation parameters, EVI was recommended to characterize the change of vegetation in this experiment. It was evident that optimum surface roughness parameters were validated to be a promising tool for soil moisture retrieval in prairie areas using Radarsat-2 data.

Understanding the dynamics of vegetation activity and its interaction with climate factors becomes an important issue in global change research. In this study, using satellite-derived normalized difference vegetation index (NDVI) data, we analyzed changes in vegetation growth at the region scale from 2000 to 2013 in Sichuan Province, China. And then explored the relationship between NDVI trends and the temperature and precipitation. The results showed: (1) The NDVI of the vegetated area in Sichuan was decreasing with a trend about −2.3 ×10−3NDVI.yr−1. About 2/3 of the vegetated area showed a negative trend of annual mean NDVI. (2) In larger areas vegetation growth was controlled by temperature. (3) Monthly NDVI was significantly correlated with the preceding month's temperature and precipitation. (4) The spatial pattern of relationship between NDVI and climatic factors was likely correlated with terrain and vegetation type.

Grassland, which has a great significance to human society, is susceptible to the drought. An efficient way to monitor the drought scale of grassland is imperative. In this paper, we proposed a new drought index named Grassland Drought Index (GDI) to monitor the drought conditions of the grassland ecosystem. The GDI was constructed by comprehensively integrating the soil moisture content, canopy water content (CWC) and precipitation together with different weights. The soil moisture with 1km spatial resolution was estimated by downscaling the AMSR-E soil moisture from 25km to 1km. The CWC was retrieved by the 1km MODIS products based on the PROSAIL model. And the precipitation was acquired by CRU data. Besides, the global distribution product of GDI in July 2010 was produced. By validation, the correlation coefficient R between the scaled GDI and the Standard Precipitation Index (SPI) was 0.6109, and the root-mean-square error (RMSE) was 0.7774.

Two-dimensional GIS are extensively applied to mineral potential mapping on a regional scale. However, these systems are unable to represent underground geological information in three spatial dimensions. The objective of this article is to overcome this defect and to develop a prototype system to qualitatively and quantitatively analyze underground mineral resources on the local scale, based on spatial data mining methods and three-dimensional GIS. The presented approach is based on 3D geological model established by the three-dimensional geological modeling software, such as Micromine, and is characterized by three-dimensional visualization, data management, and functionality for mineral resources prospectivity and quantitative assessment integrating weights-of-evidence model and case-based reasoning. The resulting system enables geologists to more accurately analyze the prospectivity and reserves of unknown mineral resource on the local scale.

A novel way to estimate the live fuel moisture content (LFMC) was explored from the ratio of canopy water content (CWC) and foliage dry biomass (FDB). The CWC was estimated using the PROSAIL (PROSPECT + SAIL) radiative transfer model from the Landsat 8 product. A weak constraint 4-D variational data assimilation method was employed to assimilate the temporally estimated leaf area index into a soil-water-atmosphere-plant (SWAP) model for optimizing the model control variables. Then, the SWAP model was reinitialized with this optimum set of control variables, and better prediction of FDB was obtained. Results showed that a high accuracy level was achieved for the estimated CWC (R2= 0.91, RMSE = 84.74 g/m2) and FDB (R2= 0.88, RMSE = 48.54 g/m2) when compared with in situ measured values. However, the accuracy level of estimated LFMC was poor (R2= 0.59, RMSE = 30.85%). Further analyses find that the estimated LFMC is reliable for low LFMC but challenged for high LFMC, which indicates that the presented method still makes sense to the assessment of wildfire risk since the wildfire generally occurs when the vegetation is in low LFMC condition.

Retrieval of vegetation parameters from remotely sensed data using a radiative transfer model is generally hampered by the ill-posed inverse problem, which dramatically decreases the precision level of retrieved parameters. The purpose of this study was to use a Bayesian network-based method to allow the alleviation of the ill-posed inverse problem. This was achieved by introducing the correlations between the model free parameters into their prior joint probability distribution (PJPD), allowing the reduction of the probabilities of unrealistic combinations. Three sampling strategies intended to design three types of PJPDs that considered different correlations (represented by a correlation matrix) were presented. They were multivariate uniform distribution composed by independent free parameters, multivariate uniform distribution based on a simple correlation matrix, and multivariate Gaussian distribution based on a complicated correlation matrix, respectively. A case study of the presented method to retrieve leaf area index (LAI) and canopy water content (CWC) using the PROSAIL_5B (PROSPECT-5 + 4SAIL) model from Landsat 8 products was implemented. Results indicate that the presented method greatly improves the precision level of target parameters, with the coefficient of determination R2of 0.69, 0.77, and 0.82 and root-mean-square error (RMSE) of 0.55, 0.51, and 0.44 m2· m-2for LAI and R2= 0.68, 0.78, and 0.84 and RMSE = 230, 198, and 166 g · m-2for CWC, respectively. Hence, the ill-posed inverse problem can be alleviated by the presented method, which can be widely applied for vegetation parameters retrieval.

The retrieval of parameters through a physical mechanism model is promising for its generality but is challenged by the ill-posed inversion problem. This study focused on the use of multiple priori information to alleviate the ill-posed inversion problem. The priori information included the products of satellite images, the correlations among model free parameters, field survey, and the achievements of previous studies. However, the priori information was of uncertainty, which was described by multi-variables probability distribution in this study. A Bayesian network algorithm was used to retrieve the canopy water content (CWC) by calculating the posterior probability distribution of CWC based on the priori information, the HJ-1B product, and the PROSAIL model. The retrieval results showed that the R2 = 0.83 and RMSE = 0.18 compared to the field measured CWC, which confirmed the feasibility to alleviate the ill-posed inversion problem by the multiple priori information.

A synergistic method of SAR and optical remote sensing data for retrieval of soil moisture was developed in this paper. Vegetation coverage, which can be easily estimated from optical data, was combined in the backscattering model. The total backscattering was divided into the amount attributed to areas covered with vegetation and that attributed to areas of bare soil. Backscattering coefficients were simulated using the established backscattering model. Then, soil moisture was estimated using the inverted model. The results showed that the predicted soil moisture correlated with the measured soil moisture (R2= 0.7075, RMSE = 3.3219 m2/m2).

Leaf area index (LAI) is one of the key vegetation indices for many biological and physical processes in plant canopies. In this study, an assimilation technique was used to simulate the LAI's varying in time series in an alpine wetland located in western China. The Terra MODIS 16 day composite surface reflectance products at 250 m resolution in 2010 with high quality were used. LAI was retrieved based on the ACRM canopy reflectance model and LUT algorithm. An experiential LOGISTIC model was fitted using the retrieved LAI, and the ensemble Kalman filter algorithm was introduced to assimilate the estimated LAI into the LOGISTIC model to update the model state.

Spatial data mining refers to extracting and “mining” the hidden, implicit, valid, novel and interesting spatial or non-spatial patterns or rules from large-amount, incomplete, noisy, fuzzy, random, and practical spatial databases. Spatial association rules mining is extracted implicit association rules from spatial database. Many metallogenic association rules lie in geology spatial database. In this paper, a method for mineral resources prediction is proposed, which mainly including uncertainty transmission between qualitative and quantitative geology spatial data using cloud model, metallogenic association rules extracting using Apriori algorithm, and comprehensive assessment of rules. At last, an experiment of iron resources prediction is performed in Eastern Kunlun Mountains, China. The results indicated that the method proposed in this paper is suitable for regional metallogenic prediction.

ASTER has higher spatial resolution in the VNIR region and higher spectral resolution in the shortwave infrared (SWIR) region in comparison with ETM+, which provide us an opportunity to distinguish various clay-alteration minerals accurately. In this paper, the 14 bands of ASTER were adopted to identify the lithologic and hydrothermal alteration minerals, such as the calcite, kaolinite and OH bearing minerals and so on, in the East Kunlun region of China. The effect of different mapping methods presented in the past, including band ratio, relative band-depth (RBD) images, False Color Composite (FCC) and SFF et al, were compared and analyzed. Results indicate that the mineral index method is suit to the high and cold mountainous area and could get the ideal result relative to the others. This work study once more again shows that the multi-spectral remote sensing techniques have excellent potentials for metallic mineral prognostication.

Dexing mines located in the east of China. They have been exploited for more 20 years and the environment pollution in this area has become very severe. It is very important to assess the spatial-temporal variation of regional environment contamination for monitoring the environment quality variation and contamination trend. 330 Sediment samples in Dexing region were collected and analyzed for As, Hg, Cd, Cr, Zn, Cu and Pb in 2004 and 1989. The maximum of As, Hg, Cd, Cr, Zn, Cu and Pb concentration in 2004 in sediments were up to 9, 4, 4.6, 1.5, 5.9, 6.3 and 5.6 times higher than theirs level in 1989, respectively. Meanwhile, the geoaccumulation index (Igeo) was used to assess the environment quality. In the end, spatial-temporal contrast was performed using GIS 3D spatial analysis from original testing data and geoaccumulation index. The contrast results indicated that there existed slightly contamination of As, Cd, Cu and Pb in 1989, mainly concentrated within little scope along the mid-lower courses of Dexing river and there were different extents of contamination of As, Cd, Zn, Cu and Pb in sediments in 2004, especially in the areas of the Dexing river, Dawu river and middle-lower courses Le’an river, while few contamination occurred in those areas in 1989. All the contrast results indicated that the extent and scope of heavy metals contamination of sediments in 2004 were bigger than that of 1989.

Leaf area index (LAI) is an important parameter for describing vegetation canopy structure in the terrestrial ecosystem on the global, continental and regional scales. In this paper, a simple data assimilation method for improving estimation of MODIS LAI time-series data products based on a new 2-D LMS (two-dimensional least mean square) adaptive filter was proposed. Firstly, The new 2-D LMS adaptive filter algorithm is introduced and analyzed. Secondly, A simple data assimilation method for improving estimation of MODIS LAI time-series data products based on the new 2-D LMS adaptive filter and quality control data of MODIS LAI is proposed. Finally, the experiments are performed based on the simple data assimilation method using MODIS LAI data products from 2000 to 2005 of southwestern China.

Crop growth monitoring is very important in agriculture resource management. Crop growth monitoring could provide crop state information for the decisive maker and reflect variety information of crop yield in time. The index of crop growth monitoring has closely relation with crop yield, which could as early as forecast large-scale food state that possibility missing or surplus. Therefore, there are important meanings to the macro control for food. Traditionally, the monitoring of crop growth and yield forecasts are made on the basis of samples by field visits or written inquiries. On national scale, the processing of these sample data is an expensive and time-consuming procedure. Recent developments in remote sensing technologies have created promising opportunities for monitoring agricultural crop growth. The moderate resolution imaging spectroradiometer (MODIS) is one detector board on Terra's (EOS-AMI), which was lunched on December 18, 1999 by NASA. It offers a unique combination of spectral, temporal, and spatial resolution compared to previous global sensors, making it a good candidate for large-scale crop growth monitoring. The paper studied the method of crop growth monitoring based on data of MODIS/TERRA vegetation indices. Results from the study not only monitor the crop growth in investigation area, but also illustrate the powerful potential to provide information about crop growth based MODIS VI data.