The remote inspection of infrastructure, like towers and dams, is crucial yet hard in remote areas with unreliable networks, and therefore depends on expensive satellite transmission. In this paper, a compression method utilizing deep learning, customized for multimodal monitoring data of infrastructure that includes both RGB and infrared inputs is proposed. The method uses a Faster R-CNN, utilizing a ResNet-50 backbone, combined with Transformer-CNN Mixture (TCM) and Transformer-Dense Mixture (TDM) blocks to pinpoint the regions of interest. The learned image compression method is then used to compress the important parts, greatly decreasing the data size for effective transfer. Experiments conducted on a custom dataset comprising 1727 pairs of RGB/infrared images demonstrate that the proposed method achieves an improvement of 0.8 dB in PSNR and 0.015 in MS-SSIM over state-of-the-art codecs while effectively preserving structural features crucial for identifying corrosion and fractures.

As a natural event that poses a serious threat to human life, property, and the natural ecology, the effective identification, assessment, and early prevention of geological hazards are crucial. Hunan Province in China is a region with a high incidence of geological hazards, exhibiting complex chain-generated characteristics due to the influence of terraced topography, heavy rainfall, and human activities. Existing landslide monitoring methods have insufficient ability to capture weak deformation at small spatial scales, making it challenging to identify landslide disaster precursors in this region effectively. This paper proposes a multi-resolution SAR collaborative monitoring method using SBAS-InSAR technology for wide-area screening, followed by a joint PS/DS-InSAR processing framework to identify weak deformation signals at small spatial scales. Using 2441 registered geohazard sites in the work area as the background dataset, wide-area InSAR monitoring and remote-sensing interpretation delineated 180 suspected geohazard target areas. Field investigation confirmed 83 of the 180 candidate target zones as active hidden-danger points, corresponding to a field-confirmed rate of 46.11% among the interpreted candidates.

Abstract Groundwater is a vital component of the hydrological cycle, and understanding its dynamics is crucial for water resource management under climate change. This study employs GRACE-FO satellite data to assess groundwater storage (GWS) dynamics in Hunan Province during the 2024 flood season (April-September). Given the abundant surface water resources in this region, we explicitly incorporate the water storage of Dongting Lake and 28 large reservoirs when calculating surface water storage anomaly (SWSA), which is crucial for estimating the GWS anomaly (GWSA). Accordingly, GWSA is obtained by subtracting the soil moisture storage anomaly (SMSA) and SWSA from the GRACE-FO-derived terrestrial water storage anomaly (TWSA). Furthermore, correlation coefficients and contribution of each water storage component to TWSA are calculated to reveal inter-component interactions and response mechanisms to precipitation. Results show that original TWSA, SWSA, and GWSA increase markedly from March to July 2024. After detrending and deseasonalizing, SWSA and GWSA exhibit a complementary relationship (correlation coefficient: −0.20), with changes of −3.08 km 3 and −1.12 km 3 over the flood season, largely attributed to anthropogenic flood control operations. In contrast, SMSA and GWSA are weakly positively correlated (0.29), reflecting limited direct recharge efficiency. TWSA is strongly correlated with both SMSA (0.78) and GWSA (0.71), reflecting synergistic variation among water storage components. Consistently, GWSA contributes the most (44.52%) to TWSA fluctuations, followed by SMSA (31.80%) and SWSA (23.68%), highlighting the critical role of groundwater in the regional water cycle. These findings provide a valuable scientific basis for sustainable water resource management and regulation in Hunan Province.

Conventional cloud removal methods often fail to fully restore details in occluded areas,thereby degrading image quality.Thus,cloud occlusion remains a persistent challenge in optical remote sensing imagery.Clouds not only obscure critical ground information but also introduce noise and artifacts during reconstruction,limiting the imagery's utility for applications such as land cover monitoring,disaster assessment,and environmental studies.Aiming to address this issue,this paper presents a cloud removal approach based on multimodal feature consistency fusion(Cloud-Harmonizer).The proposed framework leverages the complementary characteristics and consistency between Synthetic Aperture Radar(SAR)and optical imagery for the effective restoration of cloud-occluded regions and the generation of high-quality reconstructed optical images.Compared to traditional methods that rely solely on temporal or spatial interpolation,the proposed approach capitalizes on the inherent advantages of SAR data(which is unaffected by cloud cover)to support the reconstruction process and ensure the authenticity of restored areas.Through the integration of multimodal data,the method aims to improve structural and spectral recovery in cloud-affected images. The Cloud-Harmonizer framework comprises three core modules for feature extraction,alignment,and fusion of SAR and optical images.The multimodal feature consistency module maps features from the two modalities into a shared vector space and generates modality difference attention to help identify cloud-affected regions.This approach ensures compatibility between the feature representations of twomodalities,facilitating precise identification of occlusions.The consistency-constrained compensation module uses difference attention to guide SAR data in compensating for missing features in optical imagery,facilitating reconstruction that aligns with the actual scene.The multimodal collaborative adaptive fusion module adopts self-attention-based adaptive fusion strategies to optimize the integration of the two modalities and enhance overall reconstruction quality.Accurate compensation and robust feature fusion under various environmental conditions,including dense cloud coverage and complex terrain,are achieved using this modular design.The framework dynamically adjusts the fusion process using input data characteristics,increasing its suitability for diverse remote sensing scenarios. Experiments conducted on the SEN12MS-CR dataset validate the effectiveness of the proposed method.The Cloud-Harmonizer framework achieves a peak signal-to-noise ratio of 30.0408,a structural similarity index of 0.9004,and a spectral angle mapper of 7.6068,demonstrating remarkable improvements over existing cloud removal methods.These quantitative results indicate the capability of the model to recover detailed information while maintaining structural and spectral consistency in reconstructed images.Comparative analyses with existing methods indicate that the proposed approach effectively preserves textures,edges,and other details while minimizing artifacts in cloud-occluded regions.Qualitative evaluations further confirm the natural visual appearance of reconstructed images,validating the robustness of the proposed framework. Experimental results demonstrate the potential of the Cloud-Harmonizer framework for cloud removal and feature restoration in optical remote sensing imagery.Through the effective application of multimodal data fusion,the proposed method addresses cloud occlusion challenges while maintaining feature consistency between SAR and optical modalities.The approach benefits from the complementary characteristics of both data types,achieving accurate reconstruction of occluded areas while maintaining image quality.The modular and adaptive design of the framework establishes a basis for the exploration of highly sophisticated fusion strategies and the extension of applications to other remote sensing challenges.Considering the growing demand for high-quality remote sensing data,Cloud-Harmonizer may serve as a viable solution for improving the usability of optical imagery in cloud-prone environments.

The Global Navigation Satellite System (GNSS) is vital for monitoring terrestrial water storage (TWS). However, effectively extracting hydrological load deformation from GNSS observations poses a significant challenge. This study proposes a novel strategy; the seasonal hydrological load signals are removed from the raw data, and the remaining signals use principal component analysis (PCA). Simulation results from Yunnan Province demonstrate that the spatial distribution of the root mean square error (RMSE) is improved by approximately 15 % compared with traditional PCA extraction from raw data. From January 2013 to December 2022, TWS was inverted from 24 GNSS stations in Yunnan Province. The spatial distribution and time series of TWS inverted from GNSS align well with those TWS inferred from the Gravity Recovery and Climate Experiment (GRACE), GRACE Follow-On (GFO), and the Global Land Data Assimilation System (GLDAS) land surface model. However, the amplitude of the GNSS-inverted TWS is slightly higher. Since GNSS ground stations are more sensitive to hydrological load signals, they show correlations with precipitation data that are 8.6 % and 6.0 % higher than those of GRACE and GLDAS, respectively. In the power spectral density analysis of GRACE/GFO, GLDAS, and GNSS, the signal strength of GNSS is much higher than that of GRACE/GFO and GLDAS in the June and February cycles. These findings suggest that the new data extraction strategy can capture higher frequency hydrological signals in TWS, and GNSS observations can help address limitations in GRACE/GFO observations. This study demonstrates the potential of GNSS TWS in capturing higher-frequency hydrological signals and climate extremes application.

High spatial and temporal resolution remote sensing images are essential for monitoring vegetation, natural disasters, and changes in the ground surface. However, acquiring such images is challenging due to current technical limitations and cost constraints. Spatiotemporal fusion offers an effective and economical solution to achieve high spatial and temporal resolution simultaneously. This article introduces a new generative adversarial network (GAN) spatiotemporal fusion model based on multiscale convolution and attention mechanism for remote sensing images (MSCAM-GAN), to generate high-resolution fused images. The generator in MSCAM-GAN comprises three key components: feature extraction, feature fusion, and image reconstruction. Employing an encoder-decoder architecture, the generator effectively extracts multilevel features, accommodating significant resolution differences between high-resolution and low-resolution images. In the feature extraction stage, multiscale convolutional attention network (MSCAN) captures detailed features across multiple scales, dealing with spatial dependencies and long-distance relationships within the images. During the feature fusion stage, a dual parallel attention feature fusion mechanism is designed to fully integrate the extracted multiscale features. Different attention weights are assigned based on their contributions to the final output, resulting in more accurate predicted images. MSCAM-GAN was tested on the Coleambally irrigated area and lower Gwydir catchment datasets and compared with classic spatiotemporal fusion algorithms. Ablation experiments were conducted to evaluate the effectiveness of the various submodules in MSCAM-GAN. Experimental results and ablation analysis demonstrate the superior performance of the proposed method compared to other approaches.

This paper suggests a better consensus mechanism method for Delegated Proof of Stake (DPoS) in UAV-assisted mobile edge computing (MEC) for Improved Delegated Proof of Stake (I-DPoS) to solve the problems such as the ease of selecting malicious nodes as consensus nodes, the delayed handling of malicious nodes, and the tendency for certain nodes to be selected as consensus nodes without aggressive voting. Firstly, the reputation value model is proposed to calculate each node's globally unique reputation value Trust in the current election cycle T. Secondly, in the phase of proxy nodes selection, a new evaluation criterion is constructed for proxy node selection. Then, the reputation value of the node is rewarded or punished according to the node's behavior, both of the block-out node and the voting node, the reputation reward is given for successfully generating blocks, and the reputation punishment is given for malicious behavior. Finally, the experimental results show that compared to the Traditional Delegated Proof of Stake (TDPoS) and Anomaly Detection and Reward-Punishment based DPoS (ADRP) algorithms, the I-DPOS algorithm has seen a significant decrease in the proportion of proxy nodes that are maliciously elected, and at a total number of nodes of 36, the network throughput is improved by 17 transactions per second and 8 transactions per second respectively, and the average latency of the outgoing block is reduced by 3.41 s and 2.32 s respectively.

With the rapid development of mobile internet, intelligent IoT, and 5G communication technologies, many IoT devices connect to the Industrial Internet of Things, generating significant data. Blockchain is widely used in identity authentication and trust management due to its reliability. In blockchain technology, transaction packaging is a crucial component of the consensus mechanism and is critical to enhancing fairness and service quality in request processing. However, the flat structure of the blockchain, the necessity for multi-party consensus, and the profit-seeking nature of nodes lead to issues such as unfair transaction processing and prolonged response times in cloud-edge-end architectures, which are critical for empowering intelligent edge applications. To address these challenges, we have refined the blockchain consensus mechanism and introduced a novel packaging algorithm, ITFPA (Improved time-fee packaging algorithm), within the cloud-edge-end environment. This algorithm models the transaction packaging problem as a 0–1 knapsack problem and employs a branch-and-bound method to find the optimal solution. The proposed model considers both the transaction waiting time and transaction fee, using the weighted result of these factors as the priority for the transaction. We compared the proposed algorithm with the WaitTime and TxFee algorithms across four metrics: system fairness, transaction response time, and block priority. The experimental results demonstrate that the proposed algorithm enhances system fairness, reduces transaction response times, and improves service quality to a significant degree.

Our study is based on 68 continuous GNSS observation data time series in Sichuan-Yunnan covering 4 years. We detected and deleted the outliers according to the IQR law of skewness, and the world’s first Heaviside step function model of crustal strain sequence was built. In the meantime, the sequence of crustal strain was calculated, and a correlation analysis of the strain sequence of micro dynamic information was made. It was found that strain sequence is consistent with the linear trend of the tectonic movement, yet its steady state is damaged after geophysical events, e.g., earthquakes. Finally, a method of linear fitting way for a precise strain rate field based on strain time series was proposed, and the correctness of this method was verified from the algorithm and the experiment. Compared with other current commonly used methods, it was found that detecting changes of the strain slope is more sensitive than the method of only using the GNSS time series to detect changes in the slope. It was shown that the method is more accurate than the other current methods for micro dynamic strain rate field.

Synthetic Aperture Radar (SAR) is an active microwave sensor with all-day/night and all-weather detection capability, which is crucial for detecting surface water resources. Surface water-body such as rivers, lakes, reservoirs, and ponds usually appear as dark areas in SAR images. Accurate and automated extraction of these water bodies can provide valuable data for the management and strategic planning of surface water resources and effectively help prevent and control drought and flood disasters. However, most deep learning-based methods rely on manually labeled samples for model training and testing, which is inefficient and may introduce errors. To address this problem, this paper proposes a novel water-body detection method that combines optimization algorithms and deep learning techniques to automate water-body label extraction and improve the accuracy of water-body detection. First, this paper uses a swarm intelligence optimization algorithm, Dung Beetle Optimizer (DBO), to optimize the initial cluster center of the K-means clustering algorithm, which is called the DBO-K-means (DK) method. The DK method divides the training images into four categories and extracts the water bodies in them to generate the water-body labels required for deep learning model training and testing, and the whole process does not require human intervention. Then, the labels generated by DK and training data set images are fed into the Classifier–Optimizer (CO) for training. The classifier performs a dense classification task at the pixel level, resulting in an initial result image with blurred boundaries of the water body. Then, the optimizer takes this preliminary result image and the original SAR image as input, performs fine-grained optimization on the preliminary result, and finally generates a result image with a clear water-body boundary. Finally, we evaluated the accuracy of water-body detection using multiple performance indicators including ACC, precision, F1-Score, recall, and Kappa coefficient. The results show that the values of all indicators exceed 93%, which demonstrates the high accuracy and reliability of our proposed water-body detection method. Overall, this paper presents a novel DK-based approach that improves the automation and accuracy of deep learning methods for detecting water bodies in SAR images by enabling automatic sample extraction and optimization of deep learning models.

According to the glacier characteristics of remote sensing image, a context-aware deep learning semantic segmentation network for glacier extraction is proposed based on the glacier characteristics of remote sensing image. Firstly, resnet50 is introduced as the feature extraction network to achieve the accuracy and efficiency balance of glacier feature extraction. Secondly, the context-information learning of the existing semantic segmentation network is designed. The context information including the dilated-convolutional block and the max-pooled block is designed to better extract the context information of the glacier. Multiple remote sensing trained images and tested images are selected for experiment, which is qualitatively and quantitatively compared with the existing glacier feature index extraction method and other semantic segmentation network methods. The results show that the network method in the frozen lake surface, the leakage of the mountain shadow, cloud shadow and the integrity of the extraction results have a good effect, which verifies the effectiveness and robustness of the proposed method.

以3 m级TerraSAR卫星数据为例,对广西省南宁市主城区进行D-InSAR干涉测量,并结合同期土地资源全天候建设用地监测成果进行相关度分析.结果表明,主城区有两个较大的沉降漏斗,沉降趋势由城区向郊区减缓,一定程度上表明该地区的沉降并非自然地壳运动形成.进一步研究表明,高沉降漏斗区域的地下水开采及矿产开采较为严重,并且建设用地与沉降结果整体相关系数达到了0.78,建设密集区域均伴随着不同程度的沉降.根据该区域D-InSAR沉降结果,城市建设应向高稳定性、低沉降区域进行.

北京二号民用商业卫星星座由三颗高分辨率光学遥感卫星组成,是国家核准的民用商业遥感卫星,卫星自发射以来其民用应用有待开展,开展之前有必要对其进行相应测试,以测试影像本身质量及工程项目应用效果.本文以获取的湖南省常德市部分区域的北京二号原始影像为基础,研究其在国土资源部主导的土地利用变更调查遥感监测项目中的应用效果.本文采用同等分辨率的高分二号作为对比数据源,依照遥感监测项目流程对其进行遥感解译提取地类图斑,最后从图斑个数、面积、属性等方面判别北京二号应用效果.测试结果表明:北京二号卫星影像具备较高的地物识别能力以及提取面积精度,能够满足土地变更调查遥感监测1米级的监测需求.

Considering that currently the researches on the common mode error noise characteristics of longer span time series are few,this paper choose to use spatial filtering methods to analyze the longer span time series so as to establish reference framework.Focusing on the coordinate time series of 58 reference stations in Europe,and adopting the regional stacking spatial filtering method,the sectionalized stacking spatial filtering method and the coefficient weighted stacking spatial filtering method to analyze the common-mode error(CME)respectively,this research takes the coefficients of the coordinate time series after filtering and the residuals of time series as standards to make a comparatively analysis of these three spatial filtering methods.On base of this,the research estimates the common-mode error of each area and the time series before and after filtering by means of MLE(maximum likelihood estimation).The result shows that the sectionalized stacking spatial filtering method is the best for gaining the common-mode error in European areas;the common-mode error is mainly caused by flicker noise,while the white noise,random walk noise as well as other noises also make some contributions;and in the cases with relatively longer time series,the effects of common-mode errors in the N and E directions is small and can be ignored,however,the U direction is affected a lot,thus calling for consideration.

The change of urban construction land affect the subsidence directly or indirectly, the method of D-InSAR has centimeter level or even millimeter accuracy that can provide a reliable and accurate data for the research of correlation analysis of subsidence monitoring by D-InSAR and the change of urban construction land. This article takes Guiyang, Nanning city as example, using 3m level TerraSAR data to construct the Subsidence model by interferometric measurement, then compared with the Chinese national land use change remote sensing survey database at the same measure time to have a correlation analysis GIS research between subsidence and the change of urban construction land. The results shows that the integral correlation coefficient achieved 0.78 between subsidence and the change of urban construction land, the major construction area and the high density construction area are with severe land subsidence. In addition, the correlation coefficient increased from the main city to the suburbs, indicates that some of the main city causes permanent settlement and is difficult to recover. It also shows that some area subsidence caused by long-term mining or other natural factors has no strong correlation with the change of urban construction land, therefore, the results of D-InSAR subsidence monitoring have a reaction on urban construction planning, guiding urban planning to high stability, low settlement area.

选取ITRF2008框架下中国区域的10个IGS基准站2002-2014年的坐标时间序列,将它们分成3个时段:2002-2005、2006-2011、2011-2014,采用不同的噪声模型组合对它们进行噪声分析,并对结果进行对比.结果表明,最佳噪声模型在不同时段是可变的,最优噪声模型、速度场和振幅等需对应特定时段的坐标时间序列,脱离了这个时段,讨论和对比最佳噪声模型、速度场等没有意义.

针对当前主流单维度地应变分析方法难以全面反演地壳动态迁移及其趋势,无法精确反映应变与地震相关性的问题,该文提出利用多维参量解析法进行全方位地壳应变分析.利用2011-2014年GAMIT解算得到的高精度GNSS坐标时间序列经过模型化拟合得到的ITRF08框架下的川滇地区速度场,并去除欧亚板块欧拉矢量速度场进行Delaunay三角形构网,解算得到了川滇地区的主应变率场、最大剪切应变率场、面膨胀率场.研究结果表明,最大剪切应变率高值区域及面膨胀率梯度较大区域与地震的发生存在较强的相关性,地壳动态迁移趋势性体现比以往常规研究显著占优;多维参量解析法从更丰富的数据维度反演了川滇区域地壳活动状态.

In order to analyze the skewness of GPS time series data,a skewness factor MC is introduced for measurement as well as construction of the corresponding function in order to improve the detection range of the IQR law.We interpolate after rejecting the outliers,using the time series of the CHAN IGS station provided by Sopac experimental sequence,artificially adding different orders of magnitude and different densities of outlier.We confirmed that the IQR law considering skewness is more rigorous in algorithm and more effective in outlier detection than IQR law alone.

对比论证了不同欧拉矢量在小范围块体上对应变率求解的影响,并以安徽区域为例,得出了在相对小块体上求解应变率场后,由块体本身求得的自适应欧拉矢量比利用NNR-NUVEL1A提供的欧亚板块欧拉矢量更能反映其区域块体内部应变的结论。自适应欧拉矢量大大减弱了其所在块体的相对速度系统差,有效去除了其跟随大板块背景场的运动趋势,在有条件求得其自适应欧拉板块参数时,比利用大板块的欧拉矢量效果更好。