[Purposes] Soil moisture retrieval was recognized as a technical challenge limiting accurate and rapid acquisition of soil moisture data in fields such as hydrology and meteorology. [Methods] Based on the Dubois model,Sentinel-1 radar data and concurrent Sentinel-2 optical images were utilized,along with field soil moisture data obtained through the traditional drying method and time-domain reflectometry(TDR)in-situ measurements.A soil moisture retrieval model was developed for sparsely vegetated and bare soil areas,and a soil moisture map of the study area was generated by incorporating cumulative distribution function(CDF)correction results. [Findings] High consistency was observed between soil moisture data obtained by TDR in-situ measurements and the traditional drying method (R²=0.917,RMSE=0.038 7 m³·m^-3).Good agreement was also demonstrated between TDR in-situ measurement data and model-inverted soil moisture data (R²=0.706,RMSE=0.040 6 m³·m^-3).After applying the CDF method for consistency correction,the RMSE was decreased from 0.040 6 m³·m^-3 to 0.028 9 m³·m^-3,and the KS statistic was decreased from 0.173 1 to 0.086 5. [Conclusions] Data measured by TDR devices can support high spatiotemporal resolution surface moisture monitoring,and the CDF method contributes to improve data quality,providing valuable insights and references for surface soil moisture retrieval.

[Purposes] Accurately estimating glacier velocity is of great significance for the early warning of glacier-related chain disasters such as ice avalanches and debris flows. [Methods] Traditional offset-tracking methods can capture high-gradient glacier velocities,but their accuracy is limited by the image resolution.In this study,a SAR autofocusing offset tracking method is proposed,which improves the precision of offset estimation by sacrificing temporal resolution.Using medium-resolution Sentinel-1 data,the seasonal motion characteristics of the Sedongpu glacier were extracted and analyzed,and the results were validated through comparison with high-resolution Fucheng-1 data. [Findings] The results indicate that the autofocusing method effectively retrieves glacier velocities from medium-resolution SAR data,with results consistent with those from high-resolution imagery.The eastern branch of the Sedongpu glacier exhibits significantly higher summer velocities compared to winter,whereas the western branch shows no clear seasonal variations. [Conclusions] Overall,the autofocusing method demonstrates great potential for extracting mountain glacier velocities from medium-resolution SAR data.

[Purposes] To explore the application of remote sensing indices at different time sequences in soil salinity prediction and address the issues of period specificity and weak generalization ability of single-phase remote sensing data in soil salinity prediction. [Methods] The research is based on the Landsat 8/9 monthly time-series images of 2023,constructing six groups of different types of remote sensing index combinations,and generating single-phase combinations of May and September as comparisons.After screening redundant sample points using the leave-one-out method,a random forest model for predicting soil salinity is constructed.Meanwhile,the SHAP method is combined to analyze feature contributions,and principal component analysis is utilized to explore the influence of redundant sample points. [Findings] The time-series model outperforms the single-phase model.Through the feature coupling of vegetation,water and salt indices,combination 5 performs best in the time-series scenario.Time-series data integrate the full-cycle information of vegetation growth and water transfer,effectively overcoming the limitations of single-phase data.The multi-dimensional balanced feature combination can suppress the interference of redundant samples and improve the stability of the model.Different types of remote sensing indices have different responses and predictive contributions to salt content,and are closely related to the type of index. [Conclusions] The technical solution combining random forest with multi-dimensional time-series index feature coupling and redundant sample control is effective in soil salinity prediction and can provide reliable technical support for regional soil salinization monitoring.

[Purposes] Studying lake level changes of Qinghai-Xizang Plateau that known as the “Water Tower of Asia” is crucial for understanding regional ecological stability and climate change. [Methods] Here,we takes Namco Lake of Qinghai-Xizang Plateau as a case study.First,the water level of Namco Lake is extracted using SWOT satellite data.Then,the accuracy of the water level extracted by SWOT satellite is verified using DAHITI and HydroWeb altimeter datasets.Finally,we analyzed the the influence of rainfall on the water level changes of Namco Lake. [Findings] The results show that a strong correlation and consistency between the water level extracted by SWOT satellite and those from DAHITI and HydroWeb.The average RMSE values between SWOT water level and the DAHITI and HydroWeb datasets are 0.102 m and 0.054 m,respectively; while the average correlation coefficients (R²) are 0.865 and 0.942,respectively.Furthermore,the water level changes in Namco Lake show a significant positive correlation with rainfall changes,with their overall trends being largely consistent,implying that rainfall as a key factor driving the water level changes in Namco Lake. [Conclusions] This study verifies the feasibility of using SWOT satellite to monitor the water level in Namco Lake and provides a scientific basis for further revealing the relationship between lakes and climate change on the Qinghai-Xizang Plateau.

[Purposes] Aiming at the challenge of ambiguity resolution in tightly coupled GNSS and INS within complex urban environments,this paper proposes a quality control-based stepwise ambiguity fixing method. [Methods] This method classifies satellite data quality and partitions ambiguities into high-quality and low-quality subsets using the extended Kalman filter innovation sequence and IGG-Ⅲ robust estimation.A stepwise fixing strategy is adopted: the high-quality subset first undergoes decorrelation processing; after dual confirmation via the Bootstrapping success rate and fixed failure-rate ratio test,it serves as constraints to update the float solution of the low-quality subset,thereby achieving final fixation. [Findings] Experiments show that,compared with traditional methods,the proposed method improves the ambiguity fix success rate from 50.62% to 92.18%,with a valid fix rate of 91.58%.The RMS errors in the east,north,and up directions are reduced by 50.1%,48.7%,and 47.3%,respectively. [Conclusions] This method significantly enhances the ambiguity fix success rate and positioning accuracy in urban environments,providing an effective quality control solution for vehicle navigation.

[Purposes] To evaluate the impact of IGU and IGF precise ephemerides provided by the International GNSS Service on the accuracy of global navigation satellite system (GNSS)-based precipitable water vapor (PWV)retrieval. [Methods] Based on the principles of GNSS water vapor remote sensing,RTKLIB was used to process observational data from eight representative stations in the multi-GNSS experiment (MGEX)tracking network for January 2024.Using PWV from ERA5 reanalysis data as a benchmark,the retrieval results from the two ephemerides were quantitatively compared through mean bias,root mean square error (RMSE),standard deviation,and correlation coefficient. [Findings] The results indicate that the IGF ephemeris yields higher overall accuracy for PWV retrieval,with RMSE and standard deviation both at 1.9 mm,outperforming the IGU ephemeris (2.1 mm).The PWV retrieval accuracy shows significant latitudinal variation,with higher accuracy in mid-to-high latitude regions (RMSE approximately 1.4 mm)compared to regions near the equator (RMSE approximately 2.5 mm).The temporal variation of PWV is closely related to rainfall events,with rapid short-term increases (e.g.,exceeding 5 mm)often followed by rainfall. [Conclusions] This study demonstrates that the IGF ephemeris offers higher reliability in PWV retrieval.GNSS-based PWV retrieval not only reflects the spatiotemporal distribution characteristics of water vapor but also provides valuable information for short-term and imminent rainfall forecasting,offering insights to promote the operational application of GNSS water vapor remote sensing.

[Purposes] Aiming at filtering diverges caused by not-positive definite state covariance matrix of cubature Kalman filter (CKF)in practical applications and the system is unstable when there are noise anomalies,in this paper,an improved backward smoothing robust adaptive SRCKF integrated navigation algorithm is proposed. [Methods] On the basis of the SRCKF algorithm,backward smoothing is employed to improve the accuracy of filtering.Additionally,an adaptive factor has been introduced to adjust the cross-covariance matrix between the predicted state vector and the measurement vector,further mitigating the impact of observation anomalies on the system. [Findings] Through the analysis of simulation and measured data experiments,the proposed algorithm has higher filtering accuracy,and has a good effect on reducing the impact of measurement noise mutation. [Conclusions] It improves the anti-interference ability of the system,and provides reference value for the post-processing of GNSS/INS integrated navigation data.

[Purposes] This paper aims to evaluate the enhancement effect of different ionospheric correction products on the performance of smartphone single-frequency GNSS standard point positioning. [Methods] It used Huawei Mate40 and vivo iQOO smartphones as test terminals.A comparative analysis was conducted on the performance in SPP using the broadcast ephemeris-based Klobuchar model and the real-time (RT-GIM),rapid (RAP-GIM),and final (FIN-GIM)global ionospheric map products provided by the International GNSS Service. [Findings] With the Klobuchar model,the positioning RMSE for the Huawei Mate40 and vivo iQOO smartphones were 3.85 and 8.25 m,and 5.76 and 8.33 m in the north and up directions,respectively.When using RT-GIM,they decreased to 2.35 and 6.61 m (Huawei Mate40),and 3.87 and 6.45 m (vivo IQOO).The use of FIN-GIM further optimized the performance to 2.29 and 6.42 m (Huawei Mate40),and 3.47 and 5.94 m (vivo iQOO),with RAP-GIM showing similar performance. [Conclusions] The IGS ionospheric products are significantly superior to the Klobuchar model,with FIN-GIM demonstrating the best overall performance,providing an effective reference for positioning in smartphones and other single-frequency GNSS terminals.

[Purposes] This study proposes a denoising method based on the deep residual shrinkage network (DRSN)to solve the challenges of suppressing non-stationary noise in GNSS coordinate time series. [Methods] Firstly,one-dimensional convolutional neural network architecture with channel attention mechanism is const,ructed,and residual connections are utilized to avoid the gradient vanishing problem.Subsequently,the feature channel thresholds are dynamically calculated and an adaptive threshold processing mechanism is designed to achieve high-fidelity noise suppression while preserving signal details by combining global average pooling and fully connected layers. [Findings] Simulation experiments show that compared with the traditional variational modal decomposition,complementary integrated empirical modal decomposition,and deep neural network methods,the root-mean-square error of this paper's method is reduced by 41.6%,35.8%,and 15%,respectively,the signal-to-noise ratio (SNR)increases to 29.82 dB,and the correlation coefficient reaches 0.998 5.In the real-world data experiments,the method reduces the RMSE by an average of 42.3%at the three monitoring stations compared to the optimal comparative method,and the average SNR increases to 98.18 dB. [Conclusions] This proposed method significantly outperforms the other traditional methods.This study provides a new technical pathway for high-precision GNSS positioning and surface monitoring.

[Purposes] The unification of the global height datum relies on high-precision global gravity field models,yet existing models are often insufficient in both accuracy and spatial resolution. [Methods] To address this limitation,this study extends GRACE/GOCE satellite gravity field models by integrating the high-degree model EGM2008 and the residual terrain model (RTM),thereby constructing a combined gravity field model to compensate for their truncation errors.Using this combined model and GNSS/leveling data from Brazil,the vertical datum parameters for the region are determined. [Findings] Spectral analysis reveals that the satellite gravity models exhibit lower geoid degree errors and higher signal-to-noise ratios up to approximately degree 200.Validation with GNSS/leveling data shows that the accuracy of all combined models surpasses that of the individual GRACE/GOCE models and EGM2008.Among them,the GMMG2021S-EGM2008-RTM model demonstrates the optimal performance,improving accuracy by 40.7%compared to the Tongji-GMMG2021S model and by 11.3%compared to the EGM2008 model.The final calculation yields a gravity potential value of 62 636 849.27 m²·s^-2 for the Brazilian height datum.The corresponding potential difference and vertical deviation from the global datum are approximately 4.13 m²·s^-2 and 42.2 cm,respectively,indicating that the Brazilian height datum is about 42.2 cm above the global datum. [Conclusions] This research demonstrates that the spectral combination method effectively enhances model resolution and accuracy,providing a valuable reference for global height datum unification.

[Purposes] Accurate building height data are essential for characterizing urban 3D morphology,analyzing spatial structures,and formulating sustainable development strategies.This study aims to address the challenge of fine-scale building height retrieval at the regional level by constructing a high-precision inversion model for the Yangtze River delta urban agglomeration. [Methods] Building height samples are constructed based on GEDI LiDAR data.Multi-source remote sensing data from Landsat 8 and Sentinel-1/2 are integrated to extract spectral,index-based,texture,radar,and statistical features.A random forest algorithm is employed to invert building heights,and the influence of different feature combinations on inversion accuracy is evaluated. [Findings] The combination of spectral,index,texture,radar,and statistical features achieves the highest accuracy at the regional scale (R²=0.740,MAE=6.134 m,RMSE=8.651 m,MSE=74.835).The model demonstrated strong predictive performance across the Yangtze River delta (R²=0.680,MAE=5.217 m,RMSE=6.583 m,MSE=43.332),indicating good adaptability and robustness.Spatially,building height exhibits a “high core-low periphery” pattern and shows significant correlations with urban development indicators,with the total retail sales of consumer goods being most strongly related,reflecting the intrinsic law of urban spatial intensification. [Conclusions] The study confirms the effectiveness of multi-source remote sensing feature fusion for regional-scale building height inversion and highlights model performance differences across regions with varying development characteristics,providing a valuable technical reference for sustainable urban planning and 3D spatial analysis.

[Purposes] Addressing the current lack of robust watermarking algorithms suitable for remote sensing image data in multi-party collaboration scenarios,this paper proposes a secure multi-party watermarking algorithm for remote sensing image data based on secret sharing and DCT-QIM. [Methods] The algorithm first scrambles the original data in blocks,then uses additive secret sharing to divide the image into multiple encrypted shares,which are distributed to different collaborating parties.Each collaborator independently performs quantization index modulation watermark embedding on their respective shares in the discrete cosine transform domain.After collecting all watermarked shares,the watermarked image is reconstructed through additive reconstruction. [Findings] Experimental results show that the peak signal-to-noise ratio of the watermarked image reaches 50.126 dB and the structural similarity index is 0.999,indicating minimal impact on visual quality.The embedded watermark can be successfully extracted under various attacks,including cropping,rotation,Gaussian noise,and JPEG compression,demonstrating strong robustness.Furthermore,incomplete shares held by any single party cannot reveal any meaningful information about the original image,and such incomplete shares also prevent any individual party from privately verifying,tampering with,or forging watermark credentials,thereby effectively ensuring data privacy and security throughout the multi-party collaboration process. [Conclusions] The proposed algorithm enables secure collaborative multi-party watermark embedding while preserving the privacy of remote sensing image data,providing technical support for copyright protection in multi-party collaboration scenarios involving remote sensing image data.

[Purposes] Spaceborne photon-counting LiDAR systems,exemplified by ICESat-2,can directly acquire high-precision 3D information about the Earth's surface.This data serves as a novel type of control information for enhancing the 3D geolocation accuracy of satellite imagery.This study proposes a large-area block adjustment method based on topographic slope adaptive generalized iterative closest point (TSA-GICP)registration. [Methods] A strategy is designed involving global coarse registration followed by local TSA-GICP fine registration between linearly distributed spaceborne photon point clouds and digital surface models (DSMs)generated from satellite imagery.A 3D feature detection algorithm is employed to extract common feature points from the two types of heterogeneous point cloud data,enabling the automatic screening and generation of photon control points (PCPs)that are uniformly distributed,moderate in quantity,and reliable in quality.A joint block adjustment scheme incorporating an additional parameter-based rational function model (RFM)is implemented. [Findings] Compared to uncontrolled positioning and positioning assisted by Google imagery and SRTM DEM,the planimetric accuracy of ZY-3 imagery improved from 10.445 and 3.695 m to 3.058 m,and the vertical accuracy improved from 9.916 and 2.796 m to 1.821 m.Compared to Google imagery and SRTM DEM-assisted positioning,the planimetric and vertical accuracies improved by 17.24% and 34.87%,respectively.For BJ-3 imagery,planimetric accuracy improved from 6.477 and 3.879 m to 2.703 m,and vertical accuracy improved from 5.714 and 2.226 m to 0.793 m.Compared to Google imagery and SRTM DEM-assisted positioning,the planimetric and vertical accuracies improved by 30.32%and 64.38%,respectively. [Conclusions] The proposed method effectively enhances the 3D geolocation accuracy of satellite imagery,providing a feasible technical reference for improving global mapping accuracy.

[Purposes] 3D Gaussian splatting (3DGS)has emerged as a new paradigm for efficient and photorealistic 3D reconstruction.Under resource-constrained conditions,planning a limited number of supplementary viewpoints is an effective way to mitigate scene uncertainty and improve reconstruction quality.However,existing studies lack global scene guidance and often restrict uncertainty estimation to single-view observations,which limits global batch view planning and constrains its applicability in large-scale automated data acquisition and reconstruction. [Methods] To address these challenges,this paper introduces Fisher information into 3D Gaussian splatting and develops a scene-level uncertainty modeling approach that combines color and depth rendering gradients for comprehensive uncertainty estimation.A globally uncertainty-guided batch view planning strategy is then proposed to achieve efficient large-scale viewpoint optimization in a single step. [Findings] Experiments on several representative architectural scenes demonstrate that the proposed method significantly reduces data acquisition requirements while improving reconstruction completeness and efficiency,maximizing reconstruction quality with minimal resource consumption. [Conclusions] This approach enhances large-scale,high-fidelity 3D reconstruction under resource constraints and provides technical support for real-time environmental perception and digital twin applications.

[Purposes] To address the technical challenges of information loss in low-coherence areas and the difficulty in ensuring overall accuracy in time-series InSAR monitoring along lengthy railways,this paper proposes an enhanced long time-series MT-InSAR method based on an optimized interferometric combination strategy. [Methods] The design section of Shijiazhuang—Huanghua port intercity railway (under construction),which passes through the uneven subsidence funnel area of the North China Plain,was selected as a typical study area.Utilizing 91 Sentinel-1A SAR images acquired from January 2019 to January 2022,the ground subsidence rates and cumulative subsidence time series along the railway were extracted and analyzed. [Findings] It is found that significant uneven ground subsidence exists along the small mileage section of the railway,with the maximum cumulative settlement in three years reached 90 mm.The root mean square error with the levelling results is 2.02 mm,which indicates that the settlement information extracted from this paper's method is of high accuracy and reliability. [Conclusions] It also offers a solid research foundation for the safety operation of the Shijiazhuang—Huanghua port intercity railway and the ground subsidence prevention efforts in Hengshui and Cangzhou cities.

[Purposes] This paper aims to reveal the influencing factors of carbon emissions in the Yangtze River Delta region and improve the accuracy of carbon emission prediction. [Methods] This paper constructs a spatio-temporal dataset of carbon emissions in the Yangtze River Delta based on multi-source data from 2010 to 2023.The characteristics of land use change are extracted by using CLCD images,and the ARIMA-Transformer dynamic weighted hybrid prediction model is constructed to predict the carbon emission trend.The stability of the model is tested through cross-validation of the rolling window.Meanwhile,regression analysis is adopted to explore the impact of land use change and other factors on carbon emissions. [Findings] The prediction accuracy of the hybrid model is superior to that of the single model.The R² of the test set is 0.88 and the MAPE is 4.2%,which shows good stability and generalization ability.The forecast results show that the growth rate of carbon emissions in the Yangtze River Delta region has slowed down.Among them,the expansion of impermeable surfaces significantly promotes the growth of carbon emissions,while the expansion of forest land to a certain extent restrains the increase of carbon emissions. [Conclusions] Urban spatial expansion is the main driving factor for the growth of carbon emissions in the Yangtze River Delta.Optimizing the land use structure and promoting the construction of green cities are of great significance for the regional low-carbon transformation.

[Purposes] Monitoring land resource change patches based on satellite imagery is an effective measure for land resource protection.However,when predicting the area and quantity of rapidly fluctuating land resource change patches,the accuracy and stability of the existing time series prediction methods are not satisfactory. [Methods] In this paper,the prediction method for land resource change patches driven by large language models and cycle time prompts (LLM-CTP)is proposed.The LLM-CTP is the first method to apply large language models to predict land resource change patches.Meanwhile,by introducing cycle time prompts to each value in the time series,the LLM-CTP method can obtain higher prediction accuracy. [Findings] In predicting the area and quantity of land resource change patches,the comparative experiment results show that the average absolute percentage errors of the LLM-CTP method are 19.46% and 8.69% respectively,which are lower than LLMTime method (30.54%and 17.91%)and LSTM method (22.61% and 26.25%). [Conclusions] The relevant researches in this paper laid the foundation for timely warning and preventing the land resource damage from illegal activities,as well as ensuring the responsibilities of land resource protection.

[Purposes] Pole-like objects along roads are key elements that constitute the urban road skeleton and semantic scene.Their precise instance extraction is crucial for high-definition map production,autonomous driving environment perception,and smart city management.However,current extraction methods relying on a single data source struggle to achieve a balance between geometric,attributive,and semantic information.Furthermore,multi-source data fusion methods are plagued by issues such as data registration accuracy,density inconsistency,and temporal mismatch,making them unable to meet the requirements for component-level extraction of road furniture in urban scenarios. [Methods] To address the aforementioned problems,this paper proposes a method for instance extraction of road pole-like objects by fusing vehicle-borne images and point clouds.Firstly,high-precision registration is used to align image textures with point cloud geometric structures.Secondly,an initial clustering based on frustum bounding boxes generated from the point cloud is employed to separate potential pole-like objects.Then,precise extraction is performed by combining the spatial structural features of pole-like objects.Finally,principal component analysis is introduced to segment adhering or adjacent pole-like object clusters into instances. [Findings] Test results from certain road sections in Nansha district,Guangzhou,show that both recall and precision exceed 90%. [Conclusions] This method fully leverages the advantages of multi-source data,significantly improving the completeness of pole extraction and the accuracy of instance segmentation in complex urban scenes.It holds important value for enhancing the automation level of 3D understanding in road scenes.

[Purposes] As an important carrier for the digitization of three-dimensional space,traditional panoramic images often suffer from limitations such as uneven exposure,lack of completeness,loss of authenticity,stitching artifacts,and subpar image quality in complex spatial configurations,making it difficult to meet the high-precision,hyper-realistic rendering requirements in high-fidelity digitization contexts. [Methods] This paper proposes a spatial configuration-based method for differentiated acquisition and processing of 720° high-quality panoramic images under multi-spatial scales.The method firstly dynamically plans a differentiated solution for co-optical center and full-view image capture,basing on the spatial geometric features of the scene,and acquires full-frame DSLR image data through continuous multi-stage exposure,secondly gets HDR processing on the image. [Findings] Stitching image with optimizative method on single-platform can generates 720° high-quality panoramic images. [Conclusions] Experimental results demonstrate that this method overcomes the limitations of high dynamic range lighting,improves efficiency,accuracy,and image quality,and is adaptable to various scenarios,while demonstrating enhanced adaptability to diverse spatial configurations such as normal,crowded,and expansive scenarios.

[Purposes] Remote sensing extraction of building outlines serves as a vital source of foundational data for natural resource surveys and spatial planning.Addressing the challenge of low extraction accuracy,stemming from the complex geometric characteristics of building contours in remote sensing imagery,this study proposes the YOLO-CBAM model,which integrates the YOLO11 framework with the convolutional block attention module (CBAM)for enhanced remote sensing building contour extraction. [Methods] The model is trained and validated using the public GF-7 building dataset and compared against YOLOv8,YOLO11,the YOLO-CA model (incorporating the Coordinate Attention mechanism with YOLO11),U-Net,DeepLabV3,and Mask-RCNN. [Findings] Experimental results demonstrate that the YOLO-CBAM model achieves IoU,Dice index,mAP@0.5,F₁ score,maximum recall(R) values of 71.1%,81.5%,74.7%,0.72,88.0%,respectively. [Conclusions] These evaluation metrics outperform those of YOLOv8,YOLO11,YOLO-CA,U-Net,DeepLabV3,and Mask-RCNN confirming that the integration of CBAM significantly improves the accuracy of YOLO11 in remote sensing-based building contour extraction.

[Purposes] Intelligent monitoring of riverine waste and water pollution is critical for aquatic ecological protection,flood control safety,and smart water conservancy development.However,existing object detection methods suffer from low accuracy and poor robustness in complex riverine scenes due to challenges such as densely distributed small objects,strong water surface reflections,and difficulties in recognizing low-texture targets. [Methods] To address these issues,this paper proposes an enhanced model,FPN-YOLOv9,built upon YOLOv9.Specifically,a lightweight channel-spatial dual attention mechanism is innovatively integrated into the PAN-FPN multi-scale feature fusion module in the Neck layer,enabling dynamic enhancement of salient target features and suppression of background noise,thereby significantly improving detection capability for typical pollutants such as floating debris and abandoned vessels. [Findings] Experimental results show that FPN-YOLOv9 achieves a 53.3% increase in F1-score and a 23.4% improvement in accuracy,with high inference efficiency suitable for real-world deployment. [Conclusions] This study provides an efficient and practical technical solution for intelligent river patrol and dynamic water environment monitoring,while offering novel insights into refined design and optimization of feature fusion mechanisms in remote sensing object detection.

[Purposes] The Lhasa-Sangri section of the Sichuan-Xizang Railway lies within the complex tectonic zone of the Gangdise Mountains on the Qinghai-Xizang Plateau,where landslides are frequent.Accurate identification of potential landslide risk zones is essential for ensuring railway construction and operational safety. [Methods] This study focuses on a 10 km buffer zone along the Lhasa-Sangri section.Using Sentinel-1A ascending and descending orbit data from 2021 to 2023 and the SBAS-InSAR technique,surface deformation was retrieved. [Findings] The deformation rates of ascending and descending tracks in the research area are -41.33 to 15.74 mm/a and -54.45 to 13.9 mm/a,respectively.Two-dimensional decomposition was performed to obtain vertical and east-west deformation components.Combined with high-resolution optical imagery,90 potential landslide were identified.All identified zones are located in high or above landslide susceptibility areas,and the creep index α values are all greater than 0,indicating that the potential landslides are in a slow accelerating deformation stage. [Conclusions] This approach provides scientific support for forecasting,and disaster prevention of landslides along the Sichuan-Xizang Railway.

[Purposes] With the successful formation of the BeiDou-3 satellite network,the industrialization trend of BeiDou technology applications has become evident.Beyond providing location services,BeiDou also possesses short message transmission capabilities independent of operator signals.This paper aims to address the challenge of personnel safety supervision in highway tunnel construction under conditions lacking operator signal coverage. [Methods] A positioning system combining BeiDou short messages and UWB is established in a highway tunnel construction project without operator signal coverage. [Findings] Relying on the BeiDou short message receiving and sending terminal in the open area at the tunnel entrance,the system realizes the stable real-time transmission of tunnel workers' real-time locations and early warning information without operator signals. [Conclusions] It effectively ensures the safety of personnel during highway tunnel construction and improvs the intelligence and refinement level of construction safety supervision.

[Purposes] The urban-rural fringe,serving as a transitional zone with hybrid urban-rural characteristics,represents both a crucial link for integrated development and a focal area for human-land system conflicts.Its accurate demarcation is essential for precise land use management. [Methods] Using Nanchang as a case study,this study develops a multidimensional characteristic index model integrating the natural breaks method and restrictive factor approach to quantitatively identify the urban-rural fringe. [Findings] ①The model exhibited strong feasibility and efficiency in long-term identification for large cities,with outcomes aligning closely with the fringe's intrinsic characteristics.②During 2000—2024,Nanchang's urban-rural fringe evolved through “cluster emergence-rapid expansion-contiguous expansion-steady expansion-structural optimization”,expanding from 11 805.22 hm² to 54 704.37 hm² with V-shaped fluctuation in expansion rate.③The evolution characteristics of the urban-rural fringe were highly consistent with the territorial and spatial master planning of Nanchang. [Conclusions] The research results can provide new ideas and methods for defining the long-term range of urban-rural fringe.

[Purposes] This study aims to accurately assess the rooftop photovoltaic (PV)installation potential in Wuhan and support urban distributed PV planning. [Methods] This study integrates Wuhan's manually labeled dataset and public datasets for model training,systematically compares multiple deep learning semantic segmentation models to determine the optimal solution,completes rooftop classification based on urban functional zone planning,and calculates the installed capacity with availability coefficients. [Findings] The results show that the optimal model achieves a rooftop segmentation accuracy of over 96%,identifies five types of rooftop resources in Wuhan.The city's total rooftop area is 522 km²,and the rooftop PV installed capacity reaches 38 554.64 MW,presenting a pattern of “high density in core areas and large total capacity in outer suburbs”. [Conclusions] This study realizes the accurate quantification of Wuhan's rooftop PV geographic potential through the construction of a dataset adapted to local conditions in Wuhan,model selection,and quantitative evaluation of classified potential,providing a scientific basis for the layout and development of distributed PV in Wuhan.

[Purposes] To address the deficiencies of traditional geological subsidence prediction methods in nonlinear data processing and the issue of relying on experience in selecting parameters for RBF neural networks,and to achieve accurate prediction of subsidence along transmission lines. [Methods] An intelligent prediction model based on an improved particle swarm optimization algorithm to optimize the RBF neural network is proposed. [Findings] Comparative experiments using subsidence data from the Kunming transmission corridor show that the PSO-RBF model outperforms comparative models such as BP,RBF,PSO-BP,SVR,and LSTM in multiple indicators including MAE,RMSE,MAPE,and R²,with highly consistent prediction results and good generalization ability. [Conclusions] The superiority and engineering applicability of the PSO-RBF model in geological subsidence prediction are verified,providing an effective method for related intelligent early warning.

[Purposes] Seabed line detection is crucial for side-scan sonar image processing.Addressing the issues of low automation and inefficiency in existing methods,this paper aims to achieve automated and accurate extraction of seabed lines. [Methods] The traditional Canny method is improved by introducing a side-window fast guided filtering and incorporating the Perona-Malik nonlinear diffusion equation to suppress speckle noise.The Otsu algorithm is employed to achieve adaptive selection of dual thresholds,and the gradual nature and symmetry constraints of seabed lines are utilized to repair fractured edges,completing the accurate extraction of seabed lines. [Findings] Experiments results show that,in scenes with suspended matter interference,the extraction accuracy of the proposed method increases from 34% to 54% for comparative methods to 98.71%,and the Euclidean distance decreases from 72 to 85 pixels to 5.313 pixels.In scenes without significant interference,the accuracy reaches 98.27%,and the Euclidean distance is 5.382 pixels,both outperforming other comparative methods. [Conclusions] This study provides an effective solution for automated seabed line extraction and significantly enhances the robustness of extraction in the presence of suspended matter interference.