To enhance the efficiency of urban rooftop photovoltaic (PV)energy utilization and explore the application fields of real-scene 3D models,this paper proposes a method for optimizing the layout of rooftop photovoltaic panels driven by UAV (unmanned aerial vehicle)real-scene 3D models.The method employs UAV oblique photogrammetry to establish the real-scene 3D model of buildings and combines GIS spatial analysis algorithms to conduct fine-grained sunlight analysis and PV panel layout optimization.Firstly,the real-scene 3D model is optimized through selecting an appropriate level of detail (LOD)and reconstructing triangular meshes into quadrilateral meshes to improve the efficiency and accuracy of sunlight analysis.Subsequently,a parametric sunlight analysis module is developed to dynamically calculate the solar irradiance duration on the rooftop based on meteorological data,and utilizable areas are identified according to the peak sun hours.On this basis,GIS spatial analysis methods are applied to remove the influence of rooftop obstacles,demarcate the PV panel layout area,and design an optimized layout algorithm to achieve dynamic layout and visualization of rooftop PV panels.The study shows that this method can effectively improve the utilization rate of rooftop space and provide technical support for the scientific layout of building rooftop PV systems.

Addressing the bottlenecks in UAV and remote sensing-based smoke/flame detection task,such as insufficient multi-scale feature capture,complex background interference,and blurred edges,a method of smoke and flame target extraction based on the improved YOLOv12 model is proposed in this paper.The proposed method enhances multi-scale feature fusion through a mixed local-channel attention (MLCA)mechanism,improves detail retention in low-resolution images via an adaptive downsampling (ADown)module,and refines boundary regression accuracy with a customized adaptive loss function.Furthermore,by integrating the fine-tuned SAM2.1 model,the paper's method can realize pixel-level segmentation of targets within detection boxes.Experiments on the FASDD＿UAV,FASDD＿RS,and S-Firedata datasets shows that the proposed method achieves mAP50 scores of 93.1%,78.5%,and 68.2%,outperforming the baseline model YOLOv12 by 1.3%,1.5%,and 1.2%,respectively.The proposed method demonstrates significant advantages in detecting small targets,handling occluded scenarios and complex lighting conditions.Additionally,ablation experiments have confirmed the feature enhancement effects of the MLCA and Adown modules,as well as the optimization effect of Focaler-PIoU on model performance through dynamic gradient allocation.

3D building models with both geometric accuracy and realistic textures have become an important component of the national new infrastructure construction.Due to constraints such as UAV flight conditions and building layout,a large number of real texture occlusion problems occur in the texture mapping of 3D building models,which affect their visualization effects and the functions of applications such as query and measurement.Existing methods are based on a single texture image for repair and treat the occluded area as an unknown random variable,leading to possible deviations of texture repair from the real facade features of buildings.Based on the characteristic that the occlusion range of the same facade of a building varies in images from different perspectives,this paper proposes an automatic facade texture occlusion repair algorithm combining multi-view images and deep learning networks.The algorithm extracts the occluded area by using the structural similarity of multi-view textures after texture alignment,automatically synthesizes the real facade texture through the graph-cut method,and uses the DeepFill model to repair and optimize the synthesized texture.Experiments show that this method can repair the real texture of more than 40% of the occluded area,and the SSIM and PSNR values of the repaired facade texture are improved compared with existing methods.

Changes in biodiversity and their driving mechanisms are of great significance to the development of biodiversity conservation measures.Remote sensing technology has outstanding advantages in large-scale and long time-series research.In order to grasp the spatial pattern and time-series changes of animal diversity in the Yellow River basin,this study utilizes the time-series Landsat and MODIS satellite image data from 2000 to 2022,and combines the GEE,InVEST model,and GIS-related technologies to construct an index for assessing the animal diversity in the Yellow River basin based on 7 indicators in 2 aspects:species diversity and ecosystem diversity.The animal diversity assessment index of the Yellow River basin is constructed based on seven indicators of species diversity and ecosystem diversity,and the accuracy assessment is carried out based on the measured data,and the analysis of the driving factors is also carried out.The results show that:the BI constructed in this study had high precision,and the correlation with the measured species data is 0.80; the BI in the Yellow River basin shows an overall increasing trend during the 22 year perios; the area of the BI increases accounted for 70.63%of the total area,and the area of the BI decreases accounted for 29.37%; based on the analysis of the contribution of the driving factors,evapotranspiration is the primary environmental control factor.

Target localization based on moving platforms is a current research focus.However,due to limitations in sensor yaw angle errors and algorithm performance,target localization methods based on coordinate transformation still exhibit significant precision deficiencies.To address this,this paper proposes a nonlinear yaw optimization method that integrates the azimuth correction principle with coordinate transformation.Firstly,an azimuth correction model is constructed to address the inconsistency between the azimuth and yaw caused by the target being located off the principal point in the image.Next,the corrected azimuth constraint is combined with the coordinate transformation method to establish a nonlinear equation,and the Levenberg-Marquardt (LM)algorithm is used to iteratively optimize the yaw angle,effectively reducing yaw errors.Finally,Monte Carlo simulations and outdoor experiments are conducted to demonstrate the effectiveness of the algorithm.The results show that the proposed method can significantly eliminate yaw errors and achieve accurate target localization.

Aiming at the problem that the low coherence of airport runway makes it impossible for conventional InSAR technology to extract sufficient high-quality pixels,this paper uses 61 Sentinel-1A satellite data from January 2022—December 2023 to obtain the surface settlement monitoring results of Beijing Daxing International Airport by using the distributed scatterer synthetic aperture radar interferometry (DS-InSAR)method based on two-sample KS test and covariance matrix eigenvalue decomposition.The spatio-temporal distribution characteristics and the influencing factors of deformation are analyzed.The results show that the surface settlement rate in the study area ranges from -38.9～19.0mm/a,and the density and accuracy of the monitoring points obtained by DS-InSAR is higher,which can reflect more abundant surface deformation details.There are different degrees of settlement in each runway; The settlement in the West 3 flight area is relatively serious,and there are multiple sections with large settlement rate gradient values on the West 3 runway.The overall structure of the terminal building is relatively stable,and there is no obvious deformation trend.

In the current context of the booming aerospace field,starlink satellites,as large-scale non-cooperative targets,their orbit management and precise monitoring are of great significance.This paper takes thousands of starlink non-cooperative target satellites as the research object and deeply explores the effective methods for generating two-line element(TLE)for navigation using starlink ephemeris data.By comparing and analyzing the downloaded TLE data of starlink satellites,it is found that the existing TLE has deficiencies in accuracy in navigation applications.Therefore,an orbit model is established based on the least-squares method.High-precision TLE is obtained through collecting ephemeris data,defining the error function,calculating the Jacobian matrix,solving the normal equation,and iterative updating.The research clearly shows that factors such as data duration,frequency,and the time interval between the ephemeris and TLE generation moments have a significant impact on TLE accuracy.The research results have broad application prospects in fields such as satellite navigation,collision warning,and spacecraft cooperative operations,and will strongly promote the development and practical application of orbit management and navigation technologies for non-cooperative target satellites.

Unmanned vehicle navigation in complex environments (such as urban canyons or forest trails)faces challenges such as GNSS signal blockage,multipath effects,and outlier interference.Traditional EKF methods exhibit limitations in addressing these issues.Recently,factor graph optimization (FGO)has emerged as a research focus in the field of multi-sensor fusion,demonstrating superior global optimization capabilities and high accuracy.However,due to its reliance on the least-squares method,FGO lacks robustness against outliers,limiting its navigation performance in complex environments.This paper proposes a robust factor-graph-based optimization algorithm that combines Huber kernel function and chi square test for unmanned vehicle application scenarios in complex environments.The algorithm introduces wheel odometry (ODO)to assist GNSS+INS integrated navigation.Within the factor graph framework,ODO nodes are introduced as motion constraints,fusing observational data from GNSS and INS nodes.A robust kernel function is applied to enhance the algorithm's resistance to outliers.Experimental results show that the proposed algorithm achieves high accuracy and robustness in scenarios with strong multipath effects and GNSS signal outages,significantly improving navigation performance in complex environments.This provides a novel solution for high-precision unmanned vehicle navigation.

To address the limitations of existing satellite selection algorithms for smartphone GNSS pseudorange differential positioning,which include poor adaptability in complex environments and high computational complexity,this paper proposes a novel satellite selection strategy based on the code-minus-carrier observable.The method begins by identifying continuously visible satellites within a predefined time window to form an initial subset.A sliding-window averaging technique is then applied to remove the integer ambiguity component from the code-minus-carrier observations,enabling the construction of a quality evaluation metric.Satellites are subsequently screened using this metric to produce an optimized subset for positioning.Experiments were conducted in both open-sky and obstructed environments using Huawei P40 and Mate40 smartphones.Results demonstrate that in open-sky conditions,the proposed method achieves positioning accuracy comparable to that of full-constellation solutions.Under signal-obstructed conditions,it significantly improves overall positioning accuracy by 20%～25%for the P40 and 30%～47%for the Mate40,relative to the full-constellation approach.The method also exhibits enhanced stability and environmental adaptability in dynamic and challenging scenarios.With its straightforward computational process and ease of implementation,the proposed strategy offers an effective means to improve the accuracy and reliability of smartphone-based GNSS pseudorange differential positioning in complex urban environments.

Addressing the lack of systematic methodologies for vertiport site selection of electric vertical take-off and landing (eVTOL)aircraft in the context of the low-altitude economy,this paper proposes a site selection approach integrating multi-factor analysis with a bi-level optimization model.Using Chengdu as a case study,candidate sites are initially screened through overlay analysis of multi-source geographic data,including population density,land use,and traffic accessibility.A bi-level optimization model is then developed to maximize the total served population,incorporating both exponential and Gaussian distance decay functions,and solved using a hybrid strategy combining mixed-integer linear programming (MILP)and genetic algorithm (GA).By comparing two initial site selection schemes(new sites and sites incorporating existing general aviation airports),the study comprehensively evaluates theoretical served population,optimal value,and coverage rate.Results show that the normalized Gaussian distance decay model achieves better performance in terms of theoretical served population (390747 people)and population coverage rate (42.84%in the 0～100k population interval),while the exponential decay model demonstrates stronger small-area coverage capability.The final selected optimal solution is the “site selection scheme based on the normalized Gaussian distance decay model with existing general aviation airports,” which balances service capacity and coverage scope.The study reveals a nonlinear relationship between the number of vertiports and the population coverage rate,and indicates that the Gaussian decay model is more suitable for site selection in densely populated areas.The research outcomes provide a scientific basis for planning low-altitude transportation networks in Chengdu and offer a referential theoretical and practical framework for eVTOL vertiport site selection in other cities.

The electric traction system is important equipment of subway power supply facilities,mainly composed of pantographs,catenaries,traction substations and other devices.Catenary failures may cause major safety accidents.In order to quickly and accurately detect the geometric parameters of catenaries,this paper proposes a method to analyze the dynamic height and stagger of the catenary in subway tunnels based on point cloud data.This method extracts the central axis of the tunnel via the spatial projection method,cuts the point cloud according to the relative positional relationship of cross-section point clouds,and thus quickly extracts the position coordinates of the tops of the two rails and the bottom of the catenary.Through the geometrical positional relationship of these three,it can calculate the dynamic height and stagger of the catenary at any mileage section of the tunnel.Taking a subway tunnel as an example for verification,the dynamic height and stagger of the catenary extracted from point cloud data are compared with the measured values obtained by existing detection methods.Results show that the error is within the design specifications,which verifies the accuracy of the method.Based on 3D laser technology,this method can quickly,comprehensively and accurately obtain the geometric parameters of catenaries in tunnels,greatly improving detection efficiency and boasting broad application prospects.

This study aims to investigate the deformation characteristics of the annular rockfill dam in pumped-storage power plants using time-series spaceborne InSAR deformation monitoring technology.The permanent scatterer InSAR (PSInSAR)processing technique was employed,combined with high-precision digital elevation model (DEM)data,to monitor the surface deformation of the annular rockfill dam in the upper reservoir of the Zhanghewan pumped-storage power plant.The monitoring accuracy of InSAR was validated using ground-based synchronous monitoring data from high-precision measurement robots,and the deformation characteristics of the annular rockfill dam were analyzed.The results show that the dam body and slopes of the upper reservoir in the Zhanghewan power plant exhibited an overall uplift trend during the observation period,which is preliminarily attributed to the temperature increase from winter to summer.The correlation coefficient between the deformation rates of monitoring points obtained by InSAR technology and the ground-based synchronous observations reached 0.838,with a root-mean-square error (RMSE)of 7.24mm/a.The cumulative displacement of monitoring points in the annular rockfill dam is significantly correlated with temperature.The influence of temperature on the displacement of monitoring points exhibits a slow nonlinear characteristic,and the responses of different monitoring points are divergent.The displacement of monitoring points shows a strong correlation with water level changes,indicating that water levels have a significant impact on the upstream-downstream displacement of specific points.This study provides important references for the research and application of InSAR deformation monitoring for large-area structures such as annular rockfill dams.

In response to the problems of slow convergence and low accuracy in LiDAR point cloud registration in complex environments,this paper proposes a multi-stage optimization uncertainty ICP algorithm (CUBE-ICP).By using spatial distribution variance enhancement method and uncertainty regularization strategy,the accuracy and robustness of point cloud registration in complex scenes have been significantly improved.CUBE-ICP has developed a probability driven three-stage optimization framework: firstly,quantifying point cloud uncertainty based on a LiDAR error model.Secondly,capture the distribution characteristics of point clouds within the three-dimensional spatial unit through covariance matrices.Finally,the fusion of spatial distribution variance enhancement and uncertainty regularization constraints achieves a closed-loop approach from probabilistic modeling to robust optimization.The experimental results showed that the registration error of CUBE-ICP was significantly lower than mainstream algorithms such as ICP,3D-NDT,N-ICP,GICP,and LOAM in point cloud dual frame registration and continuous frame registration tasks.The CUBE-ICP algorithm proposed in this article has high performance advantages in the registration task of LiDAR point clouds,effectively solving the limitations of traditional ICP algorithms in processing complex scenes,and demonstrating stronger environmental adaptability and geometric feature adaptability.adaptability and geometric feature adaptability.

In the context of the profound integration of information technology and engineering construction,the process of tunnel construction is undergoing a gradual intelligent transformation.However,the current method of tunnel lining construction relies heavily on manual experience to adjust the tunnel lining trolley position,and the problem of over-pouring lining concrete is prominent.In light of the challenges encountered in the field of tunnel lining construction,this paper puts forward a multi-model combination method that integrates the lining model and the tunnel lining cart model.The proposed approach entails the construction of a lining pouring volume estimation model,and the execution of precise simulation and quantitative analysis of the tunnel lining construction process.The validation experiment employs the GTFS project as a case study,utilizing point cloud data from various construction periods of the tunnel lining.The experiment then simulates the lining pouring process for the designated project,integrating the design data of the lining cart.The results show that the estimation of lining pouring volume meets the demand of tunnel construction.The approach provides a foundation for decision-making,facilitating the adjustment of the lining trolley position and the estimation of the concrete pouring volume.The method effectively supports the digital simulation of tunnel lining construction processes and the rational formulation of concrete pouring plans for the tunnel lining.

Deformation monitoring of embankments is of great significance for flood prevention safety and engineering maintenance.This study employed DS-InSAR technology to monitor the deformation of the Sanshui section of the Beijiang embankment in Guangdong province,based on 74 scenes of Sentinel-1A data from 2019 to 2023.By integrating persistent scatterers (PS)and distributed scatterers (DS)signals,the spatio-temporal distribution characteristics of embankment deformation were obtained.The results indicate that: ①The annual deformation rate in the study area ranged from -9 to 6mm/a,showing significant spatial heterogeneity,with three subsidence centers formed from north to south.The maximum subsidence rate reached -9mm/a at the turning point of the embankment in the Lubao Town section (region B),while the subsidence area in the Leping Town section (region C)extended to the 200m protection zone outside the embankment.②Cumulative deformation analysis revealed a notable acceleration in subsidence during 2022—2023,with the subsidence in 2023 alone accounting for 61.5%of the total over five years.The cumulative subsidence near the Lubao Sluice reached -39mm.③Time-series analysis of characteristic points showed significant subsidence in PS1 and PS2 in September 2023,while PS3 exhibited a stable subsidence trend.PS4 underwent alternating uplift and subsidence processes.This study validates the applicability of time-series InSAR technology in monitoring earthen embankments and provides scientific support for the safety maintenance of critical embankment projects in the Pearl River Delta.

As a mega mountainous city in Southwest China,Chongqing features densely clustered high-rises and concentrated populations in its urban areas.Under the perspective of intelligent transformation in new surveying and mapping,this study addresses safety challenges posed by complex spatial environments to low-altitude economic development.This paper focus on key technologies for obstacle extraction and analysis during the compilation of 2D and 3D low-altitude visual navigation charts.For geographical environments characterized by complex terrain and densely distributed high-rises,a technical framework for obstacle extraction in visual navigation charts is developed based on 3D real-scene modeling.The achievements are applied and demonstrated in Chongqing's citywide visual aeronautical charts compilation project,supporting follow-up processes including visual flight navigation and safety early-warning systems,thereby enhancing low-altitude flight safety operations.

To address the issues of inaccuracy and low automation in traditional sub-bottom sediment classification methods based on the quality factor (Q-factor),this study utilizes variational mode decomposition (VMD)and correlation analysis to achieve accurate reconstruction of echo signals.By leveraging the prior characteristics of acoustic signals,an optimal frequency band is selected for computation,enabling robust estimation of the Q-factor.The study verifies the significance and accuracy of the Q-factor in distinguishing and characterizing different sediment types.Combined with unsupervised classification methods,it overcomes the limitations of traditional approaches in automation,achieving automatic classification of shallow sub-bottom sediments.

In order to study the high-precision,continuous,and real-time monitoring of building deformation by BDS-3,this paper conducts research based on the BDS-3 system and the CORS network.It utilizes the NTRIP protocol to transmit monitoring data in real time,employs the precise point positioning (PPP)technology to calculate coordinates in real time,and uses GAMIT/GLOBK for static differential calculation,thus realizing millimeter-level deformation monitoring.On this basis,a real-time deformation monitoring system has been developed.Experiments show that the mean value of the normalized root mean square (NRMS)for baseline calculation over 60 days is 0.1917,and the repeatability precision is better than 2mm.The weighted root mean square (WRMS)of the coordinate time series is approximately 1mm,and the static calculation reaches the millimeter level.The horizontal precision of the dynamic calculation is 5mm and the vertical precision is 13mm.It has been verified that BDS-3 system can provide solutions for the monitoring of urban buildings and has engineering application value.

Tropospheric delay is a significant error source in precise point positioning (PPP).To enhance PPP accuracy and reduce convergence time,fitting modeling of tropospheric delay in relevant regions can be performed to provide users with high-precision atmospheric parameter corrections.Addressing the limitations of traditional polynomial tropospheric delay models—which primarily consider mathematical approximation,neglect physical variations,and exhibit poor fitting performance in regions with significant elevation changes,this paper proposes an improved optimal fitting coefficient method based on a composite exponential function to model the height variation of tropospheric delay.This aims to enhance PPP performance over large areas and regions with substantial elevation variations.The results demonstrate that compared to the traditional optimal fitting coefficient model,the improved model enhances tropospheric delay accuracy across different station height intervals by 18.4%,19.5%,and 37.5%,respectively.In dynamic PPP mode,incorporating this model reduces the convergence time in the height component by an average of 20minutes.The new model demonstrably enhances both tropospheric delay accuracy and convergence speed.

Greenhouse serve as critical infrastructure in modern agriculture,where precise monitoring holds significant implications for agricultural modernization and food security.However,traditional segmentation and vectorization methods exhibit substantial limitations in high-precision boundary extraction due to complex optical characteristics and gradient transition properties of greenhouse edges.This study proposes a novel boundary extraction framework integrating gradient feature learning and geographic active contour modeling.Initially,a Vision Transformer based pretrained encoder extracts high-dimensional image features,while a multi-task segmentation decoder concurrently generates mask,edge,and gradient representations.Subsequently,a gradient field construction model guides the vectorization process,coupled with geographic active contour-based postprocessing to significantly enhance boundary smoothness and vectorization accuracy.Experimental results demonstrate superior performance over conventional vectorization methods in metrics including intersection over union (IoU)and maximum angular error,particularly excelling in complex geographic contour extraction tasks.This framework provides an innovative solution for remote sensing monitoring in facility agriculture.

To break through the bottlenecks of high dependence on samples and multi-scale feature fusion in previous remote sensing semantic segmentation,this paper proposes a remote sensing semantic segmentation method VF-UNet++,which integrates the visual foundation model SAM with the U-Net++ architecture.This method innovatively constructs a dual-stream feature collaborative modeling framework and designs a dual-stream feature interactive fusion mechanism to achieve cross-scale feature fusion.Meanwhile,aiming at the characteristics of remote sensing scenarios,it proposes an adaptation strategy of parameter freezing and domain knowledge injection,which effectively improves the generalization ability of the model under the condition of limited samples.Experiments based on the Inria Aerial Image Labeling Dataset demonstrate that VF-UNet++ outperforms the comparison models in metrics such as recall,F1-score,and mIoU.This method effectively tackles the challenges in transferring and adapting visual foundation models to the remote sensing domain,offering a reference paradigm for intelligent remote sensing interpretation under low-sample conditions.Additionally,it overcomes the limitation of inadequate fusion between global semantic features and local detailed features,achieving a dual enhancement in segmentation accuracy and model robustness within complex remote sensing scenarios.

Extreme weather events occur frequently in the Beijing-Tianjin-Hebei region,necessitating higher accuracy in atmospheric monitoring.The atmospheric weighted mean temperature (T_m) is a key parameter for GNSS-derived precipitable water vapor retrieval,but existing empirical models exhibit biases in this region.Based on data from five sounding stations in the Beijing-Tianjin-Hebei region,this study constructs a multi-factor regression model to analyze the impact of different combinations of dependent variables.The research indicates that meteorological factors (P,T)and temporal factors (DOY)have the most significant impact (with correlations of 0.83 and 0.95,respectively),while geographical factors have a smaller impact.After comparing 16models and evaluating using metrics such as root mean square error (RMSE),model 8 (P,T,DOY)is selected as the optimal model,achieving an 11.3% improvement in accuracy compared to the Bevis model,with the lowest bias,superior adaptability,and no systematic bias in residuals.This study optimizes the T_m prediction model in the Beijing-Tianjin-Hebei region,enhancing regional adaptability and accuracy.In the future,the data will be expanded and nonlinear modeling will be introduced to enhance adaptability to extreme weather events.

To address the issue of traditional contour interpolation methods failing to handle nonlinear terrain features in complex areas,a multi-layer perceptron (MLP)-based automatic interpolation method is proposed.Control points are generated by densifying contour lines,and free points with elevation constraints are created using a gridding method.Spatial coordinates are processed with positional encoding and input into an MLP model to build a nonlinear mapping from coordinates to elevations,with terrain constraints incorporated to optimize predictions.Experimental comparisons show that the method generates first contours accurately following terrain trends with smooth transitions in sharply curved areas like valleys and ridges,significantly outperforming inverse distance weighting and cubic approximation methods.The latter two exhibit node mismatches and abrupt curves in such regions.Ablation experiments confirm that elevation constraints enhance the continuity of terrain representation.The study demonstrates the effectiveness of MLP-based nonlinear modeling for contour interpolation,but it has efficiency limitations for large-scale data processing.Future research may focus on lightweight network design,constraint optimization,and distributed computing to improve practical applicability.

Strengthening ecological restoration of historical legacy mines is not only an inevitable requirement of ecological civilization construction,but also an important measure to effectively solve people's livelihood issues and protect and restore the green mountains and clear waters.This study selects the relatively concentrated and contiguous historical legacy mines in the southern hilly and mountainous area,the northern slope of the Nanling Mountains in the Hunan section,which are in urgent need of governance,as the research object.It analyzes the main ecological and environmental problems of historical legacy mines in this area,adheres to the systematic governance thinking,and takes the prominent ecological and environmental problems of the basin as the guidance to carry out engineering layout,clarify planning indicators,engineering contents,and annual plans,and analyze the expected effects to ensure the precise implementation of ecological protection and restoration projects,making the project effects quantifiable and assessable.The implementation of the research plan will restore the land resources in the study area,reshape the landform landscape,eliminate the potential threat of geological disasters,restore the water conservation function of forest vegetation,effectively promote the improvement of the stability of the ecosystem in the study area,achieve the continuous health and improvement of the regional ecological environment system.The implementation of the research plan will provide practical basis for the ecological restoration of similar historical legacy mines.

To address the limitations of existing ecological indices that often neglect water bodies and are inadequate for comprehensive regional assessment.This study develops an improved remote sensing ecological index (MRSEI)by integrated multi-source spatio-temporal data.The model incorporates five indicators—NDVI,NDBSI,TVDI,EWI,and TWI—to evaluate ecological conditions from four dimensions: greenness,dryness,wetness,and water.An empirical analysis is conducted using 2023 Landsat-8 imagery in Peixian county,Jiangsu province.The results indicate that the ecological environment in the study area had generally improved,with significant enhancements observed in three western towns and eastern lakeside regions.Compared to the traditional RSEI,the MRSEI demonstrates higher accuracy in evaluating water bodies (with a approximately 12% increase in correlation coefficient),owing to its incorporation of water-related indicators and optimized weighting.This study provides an effective tool for accurate ecological quality assessment and sustainable management at the county level,and more suitable for precise monitoring of large-scale land-water composite ecosystems.

The Qilian Mountains is located on the northeastern edge of the Qinghai-Xizang Plateau.It has always been known as the ice source reservoir and is the source of life that supports the water resources of the Hexi Corridor.Based on MODIS snow cover products,the temporal and spatial variation characteristics of snow cover and snow line in Qilian Mountains from 2011 to 2020 were analyzed,and the dominant climatic factors affecting their changes were analyzed in combination with temperature and precipitation data.Research results show that: ①The Qilian Mountains are dominated by short-day snow areas,and the annual snow areas are mainly distributed in the western part of the Qilian Mountains.②The spatial distribution of the whole year snow area and the permanent snow line is obviously different.The western section is larger than the eastern section in the basin,and the altitude shows the characteristics of normal distribution.The area of the whole year snow area in the northeast slope direction is higher than that in the southwest slope direction,and the height of the permanent snow line is lower than that in the southwest slope direction.③The annual variation of seasonal snow cover area generally shows a trend of decreasing first and then increasing.The changes of seasonal temperature and precipitation are significantly correlated with the changes of seasonal snow cover.④The influence of temperature change on snow cover in the high-altitude mountainous area of the western section is weak,and the snow accumulation caused by precipitation has a more significant impact on the annual snow cover area and the permanent snow line.The snow cover in the low altitude area of the eastern section is difficult to accumulate all year round,and the influence of solar radiation and rain erosion during the snow melting period on the annual snow cover area and the permanent snow line is more significant.

To investigate the application effectiveness of the mixture density network deep learning algorithm in nearshore water depth inversion,a portion of Haizhou Bay in northern Jiangsu province was selected as the study area.Landsat 8 satellite imagery and measured water depth data were utilized,and the accuracy of the results was compared and evaluated against those of the support vector machine model,random forest model,and fully connected neural network model.The results indicate that,overall,all four models exhibit commendable accuracy.The mixed density network model exhibits higher inversion accuracy in the experimental area compared to the other three models,with a coefficient of determination for water depth inversion of 0.86,a mean absolute error of 0.85m,and a root mean square error of 1.36m.The inversion results align more closely with the actual water depth.

The rapid iteration of new technologies in the big data era necessitates educational reform in the major of Geographic Information Science.This paper systematically analyzed and summarized the current situation and main issues of “spatial analysis” course teaching in universities of our country.By combining with the development tendency of geographic information industry in the big data era,we proposed a series of innovative reforming measures for “spatial analysis” course teaching in terms of course knowledge system,classroom teaching method,and practical teaching mode.These measures produced significant achievements,and can provide case guidance for optimizing the course education system of the Geographic Information Science major.

As a core course in the Surveying Engineering major,the teaching quality of fundamentals of geodesy directly impacts the quality of talent cultivation in this field.Based on the OBE (outcome-based education)approach,which advocates “output-oriented”,“student-centered” and “continuous improvement” teaching philosophies,and utilizing modern information technology in teaching methods,this study explores the teaching process to enhance the quality of core courses in Surveying Engineering.This research is of significant practical relevance.The paper discusses the design and implementation of a blended teaching approach across three stages: pre-class,in-class,and post-class,both online and offline.It also integrates ideological and political elements with specific teaching cases to enrich classroom content and improve teaching quality.The beneficial conclusions drawn from the implementation of blended teaching provide valuable references for teaching other courses.

To address the challenge that single-technology approaches can not simultaneously cover high-altitude blind zones and near-ground intricate components in digital preservation of ancient pagodas,this study develops an aerial-ground collaborative technical framework for fine-grained real-scene 3D modeling,enhancing the accuracy and completeness of historical building digitization.Innovatively integrating coarse-model-guided intelligent route optimization with multi-source aerial-ground data synergy: UAV photogrammetry constructs a coarse model via orthophotography to guide encircling close-range photogrammetry,enabling high-precision data acquisition of pagoda tops and facades; ground/handheld LiDAR supplements millimeter-level point clouds of base structures; ICP point cloud registration,bilateral filtering denoising,and RANSAC fusion algorithms achieve seamless integration of aerial-ground data.Applied to Luoxing pagoda,the collaborative model demonstrates significant improvements: aerial triangulation achieves a georeferencing RMSE of 0.038m,point cloud registration precision reaches 5mm,and texture ghosting is improved obviously.Structural deformations and voids at the base are fully resolved.This method overcomes technical barriers in coordinated high-altitude/near-ground data acquisition,providing an efficient,non-contact,high-precision solution for digital conservation of cultural relics,thereby robustly supporting China's national cultural digitalization strategy.

Aiming at the lack of feature classification accuracy of river features in UAV remote sensing due to spectral confusion and shadow masking,this study proposes a semantic segmentation method based on DeepLabV3+ for RGB-DSM multimodal data fusion.The method integrates the spectral information of RGB images from UAVs and the DSM elevation information generated from laser point clouds,and introduces the “spectral-structural” feature synergy mechanism to realize the pixel-level accurate recognition of typical river features,such as water bodies,embankment slopes,trees,and so on.The experimental results show that,compared with RGB,the pixel accuracy (PA)of the model is increased from 93.47% to 95.06%,and the mean intersection ratio (mIoU)is increased from 81.60% to 84.38%,with a significant increase in the IoU of the tree category under the RGB-DSM multimodal data input.The visual comparison shows that the inclusion of DSM elevation information effectively improves the problems of road boundary blurring and tree shadow misclassification,and significantly enhances the spatial structure representation of complex features.This paper demonstrates that RGB-DSM multimodal fusion has unique advantages in improving the semantic segmentation accuracy of river features.