Pixel-level annotation tasks are important in the intelligent processing of remote sensing images. For these tasks, Interactive Image Segmentation (IIS) models using click prompts are developing fast in the field of natural images. However, most interactive segmentation models using click prompts are unsuitable for remote sensing images with their current design of click prompts and their interaction schemes with image information. Based on the situation, we used a DETR-like model as the basic framework and redesigned the pixel decoder and the transformer decoder to better suit the task of IIS for remote sensing images. In the pixel decoder, we designed a click prompt with feature encoding to learn click information and a composite attention structure to facilitate interaction between click and image information, allowing the image feature at the click locations to more easily dominate annotation masks. In the transformer decoder, we utilized deformable attention, using only a single initialized query to obtain annotation masks and IoU prediction. In this paper, we trained our model on a composite remote sensing dataset and evaluated its performance on external datasets. The results showcased the model’s adaptability, achieving superior performance compared to existing methods. The code will be available at https://github.com/songbingze/ClickPromptRSIIS.

Numerous studies have been conducted on missing data recovery in remote sensing images, such as cloud removal and dead pixels restoration. Nevertheless, reconstructing continuous, extensive, and complete missing areas still poses a significant challenge. In this letter, we propose a new architecture named Extend-generative adversarial network (GAN), which leverages only a low-resolution image with relaxed requirements on spatial resolution and acquisition time as a condition to reconstruct a high-resolution image with large-scale missing areas. We equip Extend-GAN with learnable adaptive region normalization (LARN) to adjust the intensity distribution of pixels to reduce color distortion. We also introduce a new loss function into the training process of Extend-GAN, namely the structural similarity (SSIM)-based triplet loss, which helps to preserve the between missing parts and known regions. Gaofen-2 and Landsat-9 image pairs are used to validate the proposed method. Extend-GAN performs better when comprehensively evaluated on visual effect, quantitative metrics, processing speed, etc. Code is available athttps://github.com/yc-cui/Extend-GAN.

Pansharpening requires the fusion of a low-spatial-resolution multispectral (LRMS) image and a panchromatic (PAN) image with rich spatial details to obtain a high-spatial-resolution multispectral (HRMS) image. Recently, deep learning (DL)-based models have been proposed to tackle this problem and have made considerable progress. However, most existing methods rely on the conventional observation model, which treats LRMS as a blurred and downsampled version of HRMS. This observation model may lead to unsatisfactory performance and limited generalization ability at full-resolution evaluation, resulting in severe spectral and spatial distortion, as we observed that while DL-based models show significant improvement over traditional models on reduced-resolution evaluation, their performances deteriorate significantly at full resolution. In this article, we rethink the observation model and present a novel perspective from HRMS to LRMS and propose a pixel-wise ensembled masked autoencoder (PEMAE) to restore HRMS. Specifically, we consider LRMS as the result of pixel-wise masking on HRMS. Thus, LRMS can be seen as a natural input of a masked autoencoder. By ensembling the reconstruction results of multiple masking patterns, PEMAE obtains HRMS with both spectral information of LRMS and spatial details of PAN. In addition, we employ a linear cross-attention mechanism to replace the regular self-attention to reduce the computation to linear time complexity. Extensive experiments demonstrate that PEMAE outperforms state-of-the-art (SOTA) methods in terms of quantitative and visual performance at both reduced- and full-resolution evaluations. The codes are available athttps://github.com/yc-cui/PEMAE.

Due to the limitation of technology and budget, it is often difficult for sensors of a single remote sensing satellite to have both high temporal resolution and high spatial (HTHS) resolution at the same time. In this paper, we proposed a new Multi-level Feature Fusion with Generative Adversarial Network (MLFF-GAN) for generating fusion HTHS images. MLFF-GAN mainly uses U-net-like architecture and its generator is composed of three stages: feature extraction, feature fusion, and image reconstruction. In feature extraction and reconstruction stage, the generator employs the encoding and decoding structure to extract three groups of multi-level features, which can cope with the huge difference of resolution between high-resolution images and low-resolution images. In the feature fusion stage, Adaptive Instance Normalization (AdaIN) block is designed to learn the global distribution relationship between multi-temporal images, and an attention module (AM) is used to learn the local information weights for the change of small areas. The proposed MLFF-GAN was tested on two Landsat and MODIS datasets. Some state-of-the-art algorithms are comprehensively compared with MLFF-GAN. We also carried on the ablation experiment to test the effectiveness of different sub-module in MLFF-GAN. The experiment results and ablation analysis show the better performances of the proposed method when compared with other methods. The code is available at https://github.com/songbingze/MLFF-GAN.

Recently, CSR, GFZ and JPL have launched GRACE RL05 data with advantages in the spatial resolution and precision. This study explores the capability of GRACE to detect the Terrestrial Water Storage (TWS) variations in Jing-Jin-Ji Region. In this paper, we make GRACE RL05 data spatial filtering, compute and analyze TWS variations time-series and spatial trend distribution. Results show TWS in Jing-Jin-Ji Region possess obvious decreasing trend, TWS variations of CSR, JPL and GFZ from 2004 to 2014 respectively reduce in the rate of -1.13337 cm/a, -1.43606 cm/a, -1.69749 cm/a. And TWS variations spatial distribution of CSR, GFZ, JPL are strongly consistent. Besides, results also show that TWS variations accelerate significantly and decrease in southeast faster than in northwest. It is effortless to acquire that the predominant reason of TWS variations is population growth leading to badly demand for water resources. And these results tell us that TWS decreasing should be concerned.

Summary With the incessant expansion of applications and the frequent update of the software, Science Gateway for Massive Remote Sensing Image Processing (SGMRSIP), developed by client/server model or traditional browser/server model, has received more and more challenges. Fortunately, the Web 2.0 technologies, proposed in recent years, bring us a new user experience (UE) that has a fast response speed and a good interface. In particular, the remote sensing image can be processed smoothly in the absence of client software by Web 2.0 technologies. Hence, a Web 2.0‐based browser/server model is designed for SGMRSIP to enhance the UE in this paper. Firstly, functions of a parallel remote sensing image processing portal, based on high performance cluster and client/server model, are summarized. And then, a Web 2.0‐based interaction model is built, and all these functions are accomplished again on the basis of this model. Finally, the Web 2.0‐based Science Gateway is achieved. In addition, we design different workflows for different satellite data, and all the processing tasks are finished successfully to verify the feasibility of this Science Gateway. The experimental results showed that the software scalability and interaction were improved and a better UE was achieved, compared with the existing SGMRSIP. Copyright © 2013 John Wiley & Sons, Ltd.

In remote sensing applications, there are often multi-source or multi-temporal images whose different components are acquired separately. Therefore, a part of the acquired images in multi-component data can be used as priors. In this paper, the reconstruction of a remote sensing image using an auxiliary image from another sensor or another time as the reference is considered. For this application, a new compressed sensing object function with an reference image as a prior is developed. In the new model, the sparsity constraints in transform domain comes from the target image, and the gradient priors in spatial domain comes from auxiliary reference image. To optimizing the the hybrid regularization, the algorithm is based on Bregman split method. The performance of the algorithm is evaluated both qualitatively and quantitatively. The results of experiment confirm that the proposed algorithm gets higher peak signal to noise ratio (PSNR) than other approaches without reference images as priors.

Adaptive sparse representations of signals have drawn considerable interest in the past decade. In this paper, we address the problem of training dictionaries for massive images and propose a new algorithm for adapting dictionaries by extending the classical K-SVD based on only a single image. The approach presented in this paper aims at training the adapting dictionary from massive samples, other dictionary learning methods such as Online Dictionary Learning (ODL) and Recursive Least Squares Dictionary Learning Algorithm (RLS-DLA) also could train the dictionary by using relative large samples. Our method is competed with the above two state-of-the-art dictionary learning methods. Experiments demonstrate the effectiveness of the proposed dictionary learning in dealing with massive spatial-temporal remote sensing.

Ecohydrological modeling is essential to assess impact of climate change and intense human activities (land use change) on hydrological process and ecosystem to support watershed management. The traditional ecohydrological models have the deficit in coupling ecological and hydrological processes, and parameterizing vegetation parameters. Recently, optimality hypothesis, proposed by Eagleson, has been introduced to ecohydrology research which has given rise to a novel framework for modeling ecohydrological process. However, as optimality-based model has just spring up in ecohydrology, it has not been fully tested and more application of this kind of model is needed. In this study, we tried to apply an optimality-model to simulate ecohydrological process so as to test the model and support watershed management. The model has been tested in the Walnut Gulch watershed. With collected data from the study area, the model was used to simulate hourly evaportranspiration and GPP and so on. The validation result showed that, the results produced by the model were in good agreement with observed values. The VOM model can effectively overcomes the problem of traditional watershed ecohydrological models in depict ecological and hydrological coupling, the haunting task of vegetation parameters calibration. This could come to a conclusion the optimality-based ecohydrological model could be a potential approach to simulate ecohydrological process.

Marine oil spills is one of the most serious sea pollution which has a horrible effect on environment, economy, and quality of life for coastal inhabitants. How to reduce the risk of oil spill disasters has become one of the principal problems faced with marine environment management. Oil spill observation and spill processes simulation are two main parts for oil spill accident controlling and management. Traditionally, the oil spill information detection and spill simulation is disjoined without any feedback. The modeling approach is all conducted with fixed structure and static data input while the observation system is always static with fixed monitoring scheme. In such a circumstance, neither the observation system nor the simulation can provide highly accurate information. This paper propose a new framework combining oil spill monitoring and simulation as a symbiotic feedback control system based on the theory of Dynamic Data Drive Application System (DDDAS), a new paradigm dynamically integrated simulations, measurements, and applications. The numerical oil spill model can accepts real time data from remote sensing monitoring which assure modeling a more accurate and more reliable outcomes. Multiple simulations will be executed with different remote sensing monitoring scheme and the feedback from simulation guide and determine how to gather the data. For mathematical modeling of the DDDAS based marine oil spill management system, we built a multi-stage optimization model. Such system could promise more accurate prediction and more reliable outcomes with real time oil spill input, which will improve modeling technologies, advance prediction capabilities of simulation systems, and enhance oil spill monitoring.

Stream flow series analysis is of great importance for watershed management such as water conservation, water quality control etc. How to describe the characteristic of stream flow, particularly quantify stream flow variability, has been the subject of numerous studies. Due to the fact that stream flow process usually suffers from strong natural and anthropogenic disturbances, it is hardly to measure the chaotic characteristic by pure statistic method. This paper applied fractal concept for steam flow series analysis. The fractal dimension of the daily stream flow of Malian Basin was estimated using box-counting method. The result showed that the fractal dimension of the main stream and downstream gauging stations is higher than that of the tributary, headwater gauging stations. From the correlative and regress analyze of stream flow fractal dimension and soil erosion we found that the two has distinct relativity, which can bring forward a new method for the study of watershed management and so on. This could come to a conclusion stream flow series fractal dimension analysis could be a potential approach for watershed management.

In this paper, management of multi-source heterogeneous disaster background data and fast reconstruction methods of 3D scene are proposed. 2D disasters background database system and 3D disaster visualization system are designed and implemented. By the shared metadata dimensions, integrated organization of disaster background data is realized. Based on multi-threaded pre-caching technology, fast scheduling and display of remote sensing images are achieved. Through interactive mechanism, synchronized data display and operation between 2D and 3D system are realized. Finally, 3D visualization of Wenchuan earthquake area and sea level flooded simulation in the New Coastal Region of Tianjin are took to verify the feasibility of the system. The system has characters of running fast, easy to operate, simple to deploy, and suitable for practical application in disaster rescue.

As the eco-environment source region of the Yellow River, Qinghai-Tibet plateau is the startup region of climate change in China and it's very sensitive to climate change. The streamflow in this study region decrease sharply since the mid- 1980s. The streamflow shows least cycle of 5-10 years by analyzing the streamflow data in 40 years of three gauging stations (Maduo, Jimai and Tangnaihai) based on the plural Morlet wavelet analysis. The meteorological and streamflow data from 1986 to 2000 were separated into three periods, to obtain the 5 years' average data, and then the trend and cycle of the change are analyzed. The streamflow of the three stations in three periods decrease sharply compare to the average streamflow in the past 40 years, and the decreasing velocity of streamflow was increasing in recent years; especially in the Maduo gauging station, the average streamflow decrease 80% compare to the averaging streamflow in the past 40 year; the velocity of streamflow decrease from September to March in the next year is bigger than that from April to August. But the precipitation do not decrease in the three periods of three station; the daily minimum temperature increase 0.5deg, the daily maximum temperature also increase with less range; the change of surface water evaporation is irregular, which increase in the first two periods and decreased in the last period in Maduo station, while the Jimai station has opposite situation. Based on filed investigation and comparison of two periods of remote sensing data obtained in 1986 and 2000, from spatial distribution pattern variations and type transformation trend, the spatial changes and evolution pattern of land ecosystem in the eco-environment source regions of the Yellow river were analyzed using the analytical method of landscape ecological spatial pattern, the environment changed acutely. The high coveralpine cold steppe, medical cover alpine clod steppe, high cover alpine cold meadow, medial cover alpine cold meadow, swamp meadow, alpine cold swamp meadow decrease 19.35%, 24.96%, 3.32%, 9.82%, 13.08%, 40.82%, respectively; river, lake and glacier decrease 11.35%, 4.90%, 29.26%, respectively; low cover alpine cold steppe, low cover alpine cold meadow and desertification land increase 36.27 % , 3.25 % , 10.86 % , respectively. The SWAT (soil and water assessment tools) model is employed in this study, the meteorological data were obtained through "Gradient plus Inverse-Distance-Squared" (GIDS) based on the 23 weather stations in or near the study area in recent 15 years, vegetation types were derived based on the two periods of remote sensing data, the required soil parameter is obtained from the soil classification and soil texture database at the scale of 1:1,000,000. The streamflow which simulated by the SWAT model match the observed data quite well, the Nash parameter of monthly and yearly modeled streamflow in the periods of calibration and validation all exceed 0.8. After simulating the streamflow and evapotranspiration under different climate and land use scenarios, this paper try to find the impacts of climate change and LUCC on the streamflow.