Fractional snow cover (FSC) is an important parameter of snow monitoring. The snow index is generally linearly correlated with FSC; thus, the FSC estimation can be obtained by constructing a linear regression model between snow index and FSC. The conventional snow index, such as Normalized Difference Snow Index (NDSI), Normalized Difference Forest Snow Index (NDFSI), and Universal Ratio Snow Index (URSI), are usually calculated by performing ratio or difference calculations based on certain bands of the given multispectral image. In order to further improve the accuracy of FSC estimation, more useful snow features should be extracted through fully utilizing all the band information from multispectral image. Moreover, the complexity of ground object coverage brings challenges to the accurate estimation of FSC. In this letter, we proposed an effective FSC estimation method based on the deep feature snow index (DFSI). First, a Light-wideResNet model was proposed to extract DFSI by comprehensively utilizing all information from multispectral images. Then, an FSC regression model (FSCDFSI) was constructed based on the linear relationship between DFSI and FSC to estimate FSC. The proposed FSCDFSI was validated in 15 experimental areas of forest and farmland in Northeast China with a spatial resolution of 30 m. Compared with other snow index-based regression methods, the R-square (R2) of FSCDFSI could reach 0.87 and achieved the optimal accuracy with the root mean squared error (RMSE) error of FSC within 0.09.

With the rapid development of remote sensing technology, Synthetic Aperture Radar (SAR) is gradually widely used in Snow Depth (SD) estimation, and serves as a useful complement to optical sensor and passive microwave sensor for snow remote sensing applications. The selection of parameters and models related to the characteristics of snow in forest is crucial for improving the accuracy of forest SD estimation. The purpose of this letter is to develop a forest SD estimation algorithm based on optimized Feature Filtering (FF) and Deep Neural Network (DNN) using Sentinel-1 C-band data and other auxiliary data. Firstly, the optimized features are selected from input dataset using the three FF methods based on Machine Learning (ML), Maximum Mutual Information Coefficient (MMIC), and the proposed Pearson Correlation Coefficient (PCC). Then, a non-linear regression method based on DNN was developed to retrieve SD with the optimized features. Comparing results obtained from Random Forest (RF) algorithm, XGBoost (XGB) algorithm, and Recurrent Neural Network (RNN) algorithm against the meteorological stations and field measured SD data in the forests of Northeast China, the proposed method performs superior with reduced uncertainties. The Mean Absolute Error (MAE), Root Mean Square Error (RMSE), and coefficient of determination (R2) of the proposed SD estimation method are 2.98 cm, 4.30 cm and 0.77 for test dataset, respectively.

Super-resolution (SR) reconstruction of remote sensing images aims to improve image resolution while ensuring accurate spatial texture information. In most multi-scale SR methods, the feature fusion at each layer contains only the multi-scale features of the current layer. However, this approach does not optimally use these multi-scale features over different layers, leading to their gradual disappearance during the process of transmission. To address this problem, we propose a Multi-Scale Residual Dense Network (MRDN) for SR. The feature fusion of each layer in MRDN contains multi-scale features from all preceding layers, rather than only fusing the features of the current layer. Specifically, MRDN concatenates the output of each layer and passes it to the subsequent multi-scale layers to facilitate feature fusion. MRDN maximizes the utilization of hierarchical features from the original low-resolution images, enabling adaptive learning of more effective features. In addition, efficient MRDN does not necessitate a substantial increase in network depth and complexity to achieve high performance. Experimental results indicate that MRDN outperforms the state-of-the-art methods on three remote sensing datasets. To demonstrate the generalizability of MRDN, we extend its application to three relevant tasks: natural image SR, real-world image SR, and small object recognition. MRDN achieves competitive results on these tasks, confirming its generalizability.

The spaceborne scatterometer is an active nonimaging radar system that is commonly used to measure the direction and speed of wind near the ocean surface. However, the typical resolution of the current spaceborne scatterometers is 25–50 km, which limits their applicability in scenarios requiring higher resolution requirement. In this article, an adaptive bilateral-total-variation regularization algorithm with Lorentzian norm (LABTV) is proposed using HaiYang-2 Scatterometer (HY2-SCAT) data. It introduces adaptive weight coefficients based on the bilateral total variation, which suppresses the noise effectively while maintaining the texture details of the images. Moreover, the Lorentzian norm further improves the performance of the reconstruction algorithm. To investigate its performance in resolution enhancement and the resulting accuracy, the proposed reconstruction algorithm is implemented using both simulated and actual HY2-SCAT measurements. Compared with some existing resolution enhancement algorithms, the proposed algorithm can achieve a comparable resolution enhancement after enhancing two times to an original-resolution pixel size of 25 km, with the root-mean-square error (RMSE) of 1.812 dB, the peak signal-to-noise ratio (PSNR) of 26.797 dB, the structural similarity (SSIM) of 0.974, and the coefficient of determination ($R^{2}$) of 0.967, for horizontally polarized transmitted and received (HH-pol) data, and a comparable resolution enhancement with RMSE of 1.788 dB, PSNR of 26.991 dB, SSIM of 0.976, and$R^{2}$of 0.962, for vertically polarized transmitted and received (VV-pol) data. Furthermore, HY2-SCAT images with a low-resolution pixel size of 25 km were enhanced four times to a high-resolution pixel size of 6.25 km. The technique is also validated using Scatterometer Satellite-1 (SCATSAT-1) data after four times enhancement to a high-resolution pixel size of 4.45 km.

Compared with most ship detection methods for synthetic aperture radar (SAR) images, ship contour extraction can provide more of the detailed shape and edge information of an observed ship and play a significant role in sea surface monitoring and marine transportation. In this study, a joint ship contour extraction method (faster region convolutional neural network (R-CNN), fast nonlocal mean (FNLM) filter and Chan–Vese model (FFCV) method) was proposed to obtain detailed ship information from SAR images, including ship detection in complex scenes and contour extraction in target slices. First, Faster R-CNN was employed to slice ships from large-scene SAR images. Then, FNLM filtering was applied to denoise and enhance the structural information of the target slices. Finally, an optimized Chan–Vese model was proposed in this article, which can not only accurately extract the contour of the observed ship but also reduce the computation time of the model. The SAR ship detection dataset (SSDD) was selected and finely relabeled to evaluate the contour extraction performance. An evaluation index$R_{N}$, including quantitative value and offset direction, was developed to evaluate the extraction accuracy of the target contour from the SAR images. Compared with the Mask R-CNN network, the average contour extraction accuracy index$R_{N}$of the proposed FFCV method reached −0.002 on all the images in the SSDD dataset, and its results were closer to the real ship contours while maintaining the applicability to complex scenes.

Owing to the characteristics of cloud and snow, it is difficult to detect them when they coexist. Thus, we propose an end-to-end semantic segmentation network for cloud and snow detection based on hybrid feature (CSD-HFnet) to be applied to remote sensing images (RSIs) in which cloud and snow coexist. First, the local binary pattern (LBP), gray-level co-occurrence matrix (GLCM), and superpixel segmentation are combined as basic features to preserve the textural, spatial and shape information of the cloud and snow objects. We also propose deep learning feature extraction network to obtain the multi-scale deep learning features, which is utilized to better distinguish cloud from snow. The original spectral bands, the basic features, and the multi-scale deep learning features are input into the feature integration module simultaneously to form the primary multi-scale hybrid features (PMHF). Then, the PMHF are filtered, sorted and weighted by the feature filtering & sorting block and attention mechanism block to obtain the advanced multi-scale hybrid features (AMHF). Finally, the AMHF are input into the cloud and snow segmentation network, which consists of several memory-capable gate circuits for training and validation of cloud and snow detection. The results indicate that CSD-HFnet with AMHF can provide reliable detection under the condition of cloud and snow coexistence. CSD-HFnet can detect cloud and snow in multi-spectral RSIs of various spatial resolutions with an OA of 95.37%. Moreover, CSD-HFnet exhibits excellent cloud detection ability for red-green-blue (RGB) images with an OA of 95.88%, higher than other state-of-the-art cloud detection methods.

As one of the major components in the hydrological system, seasonal snow cover in Northeast China has drawn much attention recently. Because of the coarse spatial resolution of the passive microwave (PMW), heterogeneity of snowpack, and forest cover, it is difficult for existing snow products to achieve high precision snow parameters (e.g. snow depth (SD) or snow water equivalent (SWE)) assessment and hydrological research in fine scale. In this study, a novel SD retrieval algorithm that considered both the spatiotemporal dynamic of snow characteristics and forest attenuation was developed by combining the Calibrated Enhanced Resolution Brightness Temperature (TB) data and other auxiliary information, and produced a fine resolution (i.e., 6.25 km × 6.25 km) and high accuracy SD data in Northeast China. Instead of complex physical models, the machine learning was used to untangle the nonlinear complex relationship between SD and the enhanced resolution TB, forest fraction (FF), and snow characteristics. The verification results at ground weather stations showed that the retrieved SD by the proposed algorithm had high consistency with the observed SD, its RMSE, bias, and correlation coefficient (R) of 6.32 cm, -0.23 cm, and 0.63, respectively. Compared with the existing SD products (WESTDC and AMSR2), the developed model greatly improved both in spatial resolution and retrieval accuracy. In general, the fine-resolution SD inversion model achieved satisfactory accuracy and stability, and it will be used to generate long-term SD dataset service for climate change and hydrological research in the future.

Radar images with high spatial resolution are not affected by illumination or meteorological conditions, which effectively compensate for the shortcomings of optical images and passive microwave images. Thus, active microwave remote sensing technology has advantages in snow depth (SD) research. Machine learning algorithms (MLAs), which do not need to consider complex physical models, have increasingly been applied to SD research. Considering the snow conditions of different underlying surfaces, it is very important to select the appropriate parameter combinations (PC) reflecting the SD information for MLAs. In this study, C-band SAR data with 20 m spatial resolution, ground-based SD observation data from meteorological stations, and field measurement data in Northeast China were used to construct and validate the SD estimation method. Two parameter selection methods including the correlation coefficient method and the machine learning (ML) fusion method were proposed to discuss the influence of different PC on SD estimation. Then, XGBoost, random forest (RF), linear support vector regression (LSVR), and kernel support vector regression (KSVR) were applied to estimate SD based on the selected PC, and evaluate the accuracies of SD estimation using different MLAs. The results demonstrated that the PC selected using the correlation coefficient method and XGBoost algorithm could achieve the best SD results in the study area. Combining RPC-C with XGBoost algorithm, in cropland areas, the average values of mean absolute error (MAE) and root mean squared error (RMSE) were 1.75 and 2.58 cm, respectively. Combining RPC-F with XGBoost algorithm, in forest areas, the average values of MAE and RMSE were 3.12 and 5.07 cm, respectively. The research of this letter can select the optimal PC for MLAs and effectively improve the accuracy of SD estimation using C-band SAR data.

Considering the tradeoff between spatial resolution and temporal resolution, spatiotemporal fusion has become a promising technique to monitor snow cover dynamics with both high spatial and temporal resolutions. The representative spatiotemporal fusion methods, e.g. Spatial Temporal Data Fusion Approach (STDFA), usually exist obvious phenomenon of spectral distortion when the surface reflectance changes nonlinearly, which affects the quality of the spatiotemporal fusion image. To address this issue, an effective STDFA-Matching-Pix2pix-Generative Adversarial Network (SMPG) algorithm combining the unmixing-based method, deep learning method, pre-matching and post-matching module is proposed to reduce the spectral distortion of STDFA fusion image. The high-temporal-low-spatial (HTLS) resolution MOD09GA data and high-spatial-low-temporal resolution (HSLT) Landsat 8 data are selected in this study. SMPG algorithm is firstly employed to obtain daily high-spatial-high-temporal (HSHT) images, and then daily snow cover results with a spatial resolution of 30 m are obtained by calculating the normalized difference snow index (NDSI). SMPG algorithm is further compared with STDFA, Spatial and Temporal Adaptive Reflectance Fusion Model (STARFM), Flexible Spatiotemporal DAta Fusion (FSDAF), Swin SpatioTemporal Fusion Model (SwinSTFM), and Generative Adversarial Network-based SpatioTemporal Fusion Model (GAN-STFM). The experimental results indicate that the proposed algorithm yields better overall performance in daily spatiotemporal fusion image and snow cover result with a spatial resolution of 30 m. The mean correlation coefficient (CC) of SMPG can achieve 0.962, which is 0.06-0.36 higher than that of other spatiotemporal fusion methods. The error between the percentage of snow cover area obtained through SMPG and validation data is within 0.84%.

Accurate and timely crop mapping is essential for agricultural applications, and deep-learning methods have been applied on a range of remotely sensed data sources to classify crops. In this article, we develop a novel crop classification method based on spatiotemporal deep-learning fusion technology. However, for crop mapping, the selection and labeling of training samples is expensive and time consuming. Therefore, we propose a fully automated training-sample-selection method. First, we design the method according to image processing algorithms and the concept of a sliding window. Second, we develop the Geo-3D convolutional neural network (CNN) and Geo-Conv1D for crop classification using time-series Sentinel-2 imagery. Specifically, we integrate geographic information of crops into the structure of deep-learning networks. Finally, we apply an active learning strategy to integrate the classification advantages of Geo-3D CNN and Geo-Conv1D. Experiments conducted in Northeast China show that the proposed sampling method can reliably provide and label a large number of samples and achieve satisfactory results for different deep-learning networks. Based on the automatic selection and labeling of training samples, the crop classification method based on spatiotemporal deep-learning fusion technology can achieve the highest overall accuracy (OA) with approximately 92.50% as compared with Geo-Conv1D (91.89%) and Geo-3D CNN (91.27%) in the three study areas, indicating that the proposed method is effective and efficient in multi-temporal crop classification.

Inspired by recent promising developments by deep learning networks, this letter presents a novel approach based on a multifeature long short-term memory (LSTM) network with an optimal unit number for extracting buildings from multitemporal high-resolution optical satellite imagery. The algorithm first extracts specified multifeature data of buildings, including spectral, shape, texture, and indices from multitemporal high-resolution imagery. Next, a designed LSTM network architecture with an optimal unit number is trained using the multifeature data to extract buildings at the pixel level. The proposed network is compared with popular deep learning-based networks, including VGG, U-Net, and ResNet. Finally, postprocessing (e.g., morphological algorithm) is employed to optimize the classification results produced by deep learning. The approach is validated on a newly created data set using Gaofen-2 (GF-2) imagery that contains various buildings on a campus with different materials, shapes, sizes, heights, and directions. Experimental results indicate that the proposed approach outperforms current deep learning-based methods for building extraction and has a potential for practical applications.

Soil Moisture plays an important role in regulating rainfall infiltration and surface evaporation. In this study, the monthly mean data (soil moisture, sunshine percentage, mean air temperature, mean wind speed, mean relative humidity, precipitation, and mean water vapor pressure) from 1961 to 2010 were provided by China Meteorological Data Network and the soil moisture products of four layers (0-10cm, 10-40cm, 40-100cm and 100-200cm) for 1948-2010 years were provided by GLDAS (Global Land Data Assimilation System). Based on these data, the error characteristics of GLDAS soil moisture products are evaluated. Crop water requirement is calculated based on the crop coefficient method by Penman-Monteith formula. The results showed that the soil water content of 0-200cm in the northeast region decreased in recent 60 years, and the whole northeast region decreased about 92 billion m3. Combined with the water demand of maize growing season, it can be seen that the water deficit boundary of maize planting moved eastward about 1.5°. These results are of great significance for agricultural management in Northeast China in the future.

During the saline-alkali land improvement process, the soil conductivities throughout the whole land area are usually investigated in advance to grade the soil salinization. As a new soil investigating technique, ground penetrating radar (GPR) has been widely used in geographical and agricultural applications. Nevertheless, there are still challenges in applying GPR to saline-alkali soil conductivity predictions, because it is very difficult to separate the direct path wave (DPW) and the land surface reflecting echoes under near-field condition. To realize fast soil conductivity predicting, a novel waveform correlation analyzing method based on the pulsed GPR technique is proposed in this article. By this method, the saline-alkali soil conductivities can be estimated directly by waveform comparison instead of complex equation solving. Experiments are performed on specific saline-alkali land with different surface morphologies located in the western Jilin province, northeast China. An empirical relationship between GPR echoes and the saline-alkali soil conductivities is established by the measurements in the first survey line. Then, this empirical formula is used to predict the soil conductivities in the second survey line to verify the methodology. Analyzing results show that the estimated conductivities are consistent with the WET (water content, electrical conductivity, and temperature) sensor measurements and the soil sample measuring results. Based on the nondestructive GPR technique, the proposed method can provide a fast and efficient way to estimate the conductivities of the saline-alkali land.

The planting pattern (e.g. row orientation, width and spacing) has an important effect on the photosynthesis and yields of maize. UAV remote sensing offers a potential way to extract these pattern parameters conveniently and fast. A method based on the imagery acquired with an unmanned aerial vehicle (UAV) is proposed to extract row-structure parameters of the maize. A revised parallel-beam radon transform (RPRT) is performed and row orientation angle, row width and space distance are determined in Radon transform domain. The validation results using the in-situ measurements data show that the extraction results have high accuracy. These parameters obtained by UAV imageries can be as the input ones to maize growth and yield estimation model and later used in decision support tools.

Several spectral unmixing techniques using multiple endmembers for each class have been developed. Although they can address within-class spectral variability, their unmixing results may have low unmixing resolution when the within-class variation is large due to the associated high uncertainty. Therefore, it is critical to represent data in an effective feature space so that the endmember classes are compact with small variation. In this letter, a minimum-class-variance support vector machine (MCVSVM) is further developed to extend its functions for both classification and spectral unmixing. Moreover, analytical expressions for spectral unmixing resolution (SUR) are provided to measure the spectral unmixing uncertainty in the new feature space. The extended MCVSVM (e_MCVSVM) can improve SUR and reduce the spectral unmixing uncertainty as it can effectively maximize the between-class scatter while minimizing the within-class scatter. Experimental results show that the e_MCVSVM algorithm performs better in terms of the unmixing accuracy and the computation speed compared with the other algorithms (e.g., fully constrained least squares and endmember bundles) in both linearly separable and nonseparable cases. This newly proposed approach advances the linear spectral mixture analysis with greater speed and higher accuracy based on the SVM after the SUR is effectively characterized.

Due to their low spatial resolution (≥ 10 km), passive microwave data from satellites often contain many mixed pixels. This is one of the main reasons that the retrieval accuracy of snow depth data is unsatisfactory for current practical demands. This paper proposes a multifrequency dual-polarized passive microwave unmixing method for snow. The land cover type of the observation area is partitioned into broadleaf forest, needleleaf forest, annual grass vegetation, and urban types. The component brightness temperature (CBT) of each land cover type, that is, unmixed data, with a 500-m data resolution, is obtained using the proposed unmixing method. Considering that the actual snow grain size was larger than 0.4 mm in the study area, Foster's snow depth retrieval algorithm, which refines the empirical coefficient, is employed in this paper. Compared with Chang's snow depth retrieval algorithm, the overall accuracy of snow depth data improved by approximately 29.2% using Foster's algorithm. Based on Foster's algorithm, the obtained results indicate that the overall accuracy of the snow depth data improved by approximately 22.9% when using the CBT compared with the original mixed-pixel method. The accuracy of snow depth data in annual grass vegetation is improved by approximately 38%, whereas those in needleleaf forest and broadleaf forest are improved by approximately 4.6% and 4.3%, respectively. The experimental results demonstrate that the CBT effectively improves the retrieval accuracy of snow depth data when compared with the case of using only the mixed-pixel method.

Due to the limited spatial resolution of multispectral/hyperspectral data, mixed pixels widely exist and various spectral unmixing techniques have been developed for information extraction at the subpixel level in recent years. One of the challenging problems in spectral mixture analysis is how to model the data of a primary class. Given that the within-class spectral variability (WSV) is inevitable, it is more realistic to associate a group of representative spectra with a pure class. The unmixing method using the extended support vector machines (eSVMs) has handled this problem effectively. However, it has simplified WSV in the mixed cases. In this paper, a further development of eSVMs is presented to address two problems in multiple-endmember spectral mixture analysis: 1) one mixed pixel may be unmixed into different fractions (model overlap); and 2) one fraction may correspond to a group of mixed pixels (fraction overlap). Then, spectral unmixing resolution (SUR) is introduced to characterize how finely the mixture in a mixed pixel can be quantified. The quantitative relationship between SUR and WSV of endmembers is derived via a geometry analysis in support vector machine feature space. Thus, the possible SUR can be estimated when multiple endmembers for each class are given. Moreover, if the requirement of SUR is fixed, the acceptance level of WSV is then limited, which can be used as a guide to remove outliers and purify endmembers for each primary class. Experiments are presented to illustrate model and fraction overlap problems and the application of SUR in uncertainty analysis of spectral unmixing.

Extended support vector machines (ESVM) was introduced recently for spectral unmixing. It models a class using a group of representative spectra to accommodate within class spectral variation. This paper presents a further geometry analysis of this method, and an improved ESVM is developed, which takes into account both within-class spectral variability and within each mixed case. The experiments illustrate that the new proposed algorithm can obtain more realistic unmixing results.