A high seven satellite water three-dimensional mapping and three-dimensional modeling method and system

Abstract

The present application relates to the field of space photogrammetry and water conservancy surveying and mapping technology, and particularly relates to a Gao Fen 7 satellite water conservancy stereoscopic mapping and three-dimensional modeling method and system, comprising: using a water conservancy facility topology characteristic quantitative evaluation model to screen the same-track stereo images, and performing same-track stereo image splicing and fusion based on river linear topology, reservoir area surface topology and canal system node topology; dividing terrain subareas and establishing a terrain slope and RPC parameter correlation correction model. The present application takes water conservancy facility topology characteristics as the core driving force, runs through the whole process of image preprocessing, space three solution, image control layout, point cloud processing and three-dimensional modeling, replaces the precision control role of traditional dense image control points through topology constraint, solves the problems of insufficient image control precision, high image control point layout cost and low large scene modeling efficiency of existing satellite stereoscopic mapping in complex water conservancy topography, and realizes efficient and high-precision application of Gao Fen 7 satellite images in water conservancy engineering surveying and mapping.

Description

Technical Field

[0001] This invention relates to the field of aerospace photogrammetry and water conservancy mapping technology, specifically a method and system for three-dimensional mapping and modeling of water conservancy using the Gaofen-7 satellite. Background Technology

[0002] The Gaofen-7 satellite can be used for 1:10,000 scale stereo mapping and is a core data carrier for spaceborne photogrammetry. Water conservancy engineering surveys mainly employ two techniques: satellite imagery stereo mapping and UAV oblique photogrammetry modeling. Satellite imagery stereo mapping first acquires image data, then establishes image control points, performs RPC aerial triangulation and joint adjustment, and finally generates digital elevation models, orthophotos, and digital line maps to achieve topographic mapping. UAV oblique photogrammetry modeling acquires small-scale oblique images, processes them through preprocessing and modeling software to generate 3D models, and is mostly used for small-scale, high-precision modeling; it is a commonly used water conservancy surveying modeling method in the industry.

[0003] Current water conservancy engineering surveying and modeling suffers from the following problems: existing satellite stereo mapping has not been validated for accuracy under complex terrain conditions without image control points (ARPs); ARP deployment has not been optimized in conjunction with the imaging characteristics of the Gaofen-7 satellite and the distribution patterns of water conservancy facilities; and the coverage and processing efficiency of UAV modeling are not suitable for large-scale water conservancy operations. These issues directly lead to uncertain mapping accuracy of Gaofen-7 satellite imagery in water conservancy projects, high cost and risks in deploying ARPs in remote areas, and difficulties in efficiently advancing the modeling of large-scale water conservancy projects such as large reservoirs and long-distance water conveyance lines. The surveying advantages of the Gaofen-7 satellite cannot be fully realized in water conservancy projects, affecting the accuracy and efficiency of water conservancy surveying and modeling. Summary of the Invention

[0004] To address the shortcomings of existing technologies, this invention provides a method and system for three-dimensional mapping and modeling of water conservancy using the Gaofen-7 satellite, which solves the problems of insufficient accuracy in existing satellite water conservancy mapping due to the lack of image control points, unreasonable layout of image control points, and low efficiency in large-scene modeling.

[0005] To achieve the above objectives, the present invention provides the following technical solution: a method for three-dimensional mapping and modeling of water conservancy systems using the Gaofen-7 satellite, comprising:

[0006] A quantitative evaluation model of the topological characteristics of water conservancy facilities was used to screen the same-track stereo images, and the same-track stereo images were stitched and fused based on the linear topology of the river channel, the planar topology of the reservoir area, and the node topology of the canal system.

[0007] The terrain is divided into zones and a correction model is established to correlate terrain slope with RPC parameters. The rigid topological features of hydraulic engineering are extracted to construct constraint equations to correct the original RPC parameters and perform error compensation.

[0008] A virtual constraint-driven, imageless, three-solution framework for hydraulic geometric features is constructed, and a multi-scale topological feature matching algorithm is used for adjustment iterative solution and accuracy self-calibration.

[0009] Establish a topology weighting system for water conservancy facilities, and guide the sparse deployment of image control points and perform joint adjustment optimization through topology weights;

[0010] Extract the 3D point cloud from the stereo image and perform adaptive filtering. Then, perform digital elevation model revision for the feature regions to enhance elevation accuracy.

[0011] Based on water conservancy semantic segmentation, 3D modeling and cross-scene topology alignment are performed, and full-element topology verification and semantic information embedding are performed on the model.

[0012] Furthermore, the method of using the quantitative evaluation model of water conservancy facility topology features to screen the same-track stereo images includes: setting a standard threshold for the overlap of stereo models, limiting the cloud coverage area of ​​a single scene image, comprehensively evaluating image contrast, noise level and imaging angle indicators to complete the initial screening of images, and extracting regional segments with complete water conservancy topology features from the initial screened images.

[0013] The same-track stereo image stitching and fusion includes stitching cloudless areas based on hydraulic topology feature matching, generating a complete hydraulic topology fused image through grayscale normalization processing, and establishing a feature association index between supporting parameters and images.

[0014] Furthermore, the division of terrain zones includes: identifying terrain zones such as high-altitude water conveyance, hilly river channels, and plain irrigation areas;

[0015] The correction of the original RPC parameters includes: extracting the river centerline, dam outline, and canal nodes as rigid topological features of the water conservancy system; constructing a correlation correction model between the terrain slope of each region and the RPC parameters; performing regional iterative correction of the original RPC parameters; eliminating satellite imaging system bias through geometric constraint adjustment iteration; and achieving the adaptation of RPC parameters to the topological features of the water conservancy scene.

[0016] Furthermore, the imageless spatial tri-solution framework introduces terrain constraints, hydraulic topology constraints, and publicly available terrain data constraints.

[0017] The multi-scale topological feature matching algorithm performs feature fusion on water surface, slope and tidal flat areas to improve the stability and number of matching points.

[0018] The adjustment iterative solution uses virtual features of hydraulic topology as constraints in the calculation, eliminating the need for ground control points to complete aerial triangulation.

[0019] Furthermore, the accuracy self-calibration includes: collecting measured data of water conservancy points, comparing the measured data with the solution results point by point, constructing an error feedback mechanism to adjust the solution parameters, and forming a standard for judging the accuracy of water conservancy imageless control mapping.

[0020] Furthermore, the topological weighting system includes assigning high weight values ​​to dam nodes, canal intersections, and river bends;

[0021] The sparse deployment of image control points includes: calculating the deployment density of image control points in different weighted regions by combining the imaging resolution of Gaofen-7 satellite and stereo baseline parameters, and using hydraulic geometric continuity to achieve full-domain accuracy transfer;

[0022] The adjustment and joint optimization includes: jointly adjusting the virtual feature points of hydraulic geometry with the real control points, and introducing the accuracy loss coefficient and operation efficiency coefficient as evaluation indicators to determine the sparse layout scheme.

[0023] Furthermore, the adaptive filtering process includes: distinguishing between feature points and ground points based on hydraulic topological features, automatically eliminating non-topographic interference points and retaining hydraulic topographic feature points, and adjusting the point cloud thinning density according to different topographic zones;

[0024] The digital elevation model revision includes: performing geometrically guided elevation revision on river cross sections, dam foundations, and canal slopes to ensure that the digital elevation model meets the geometric requirements of water conservancy surveying and mapping.

[0025] Furthermore, the three-dimensional modeling includes: identifying water conservancy facilities based on water conservancy semantic segmentation and constructing semantic association rules, and fusing digital elevation models with orthophotos to generate a full-domain three-dimensional model;

[0026] The cross-scene topology alignment includes: using a hydraulic topology centerline alignment algorithm to process the stitching area of ​​multi-scene image models, eliminating stitching gaps and accuracy deviations;

[0027] The full-element topology verification includes: checking the compliance of the spatial relationships, geometric dimensions, and geometric connections of water conservancy facilities.

[0028] Furthermore, the semantic information embedding includes: writing water conservancy semantic information and engineering attribute data into the model, and establishing a mapping relationship between geometric elements and water conservancy business data;

[0029] It also includes lightweight geometric optimization of the model and adjustment of the model data format and accuracy indicators to adapt to the access standards of the water conservancy digital management platform.

[0030] This invention also provides a Gaofen-7 satellite-based water conservancy stereo mapping and 3D modeling system, comprising:

[0031] The topology feature preprocessing module is used for hydraulic topology evaluation, image fusion, coordinate transformation, and RPC parameter fine-tuning, and outputs standardized topology mapping data.

[0032] The topology-constrained image control-free solution module is used to execute the aerial triangulation iterative algorithm and accuracy self-calibration, and generate mapping results reports;

[0033] The topology weighting and sparse control point deployment module is used to calculate topology weights and generate sparse control point deployment schemes, and output field operation guidance documents.

[0034] The topology filtering and modeling module is used to perform point cloud topology filtering, model revision, semantic modeling, and cross-scene alignment.

[0035] The topology verification and output module is used to perform model topology verification, semantic embedding, lightweight optimization, and format conversion, and output the final results that conform to the water conservancy digital platform.

[0036] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0037] This invention utilizes the topological features of water conservancy facilities to guide the entire process of water conservancy stereo mapping and 3D modeling. It replaces traditional dense image control points with topological constraints to achieve accuracy control, verifying the accuracy of image control-free mapping of Gaofen-7 satellite imagery in complex water conservancy terrains. By combining satellite imaging characteristics with the distribution patterns of water conservancy facilities, it optimizes the layout of image control points, significantly reducing the cost and operational risks of setting up image control points in remote areas. Simultaneously, it improves the modeling coverage and processing efficiency of large-scale water conservancy scenarios such as large reservoirs and long-distance water conveyance lines, fully leveraging the mapping capabilities of Gaofen-7 satellite in water conservancy projects. This effectively improves the accuracy and operational efficiency of water conservancy engineering mapping and modeling, and also establishes an accuracy judgment standard for image control-free mapping adapted to water conservancy scenarios. The optimized digital elevation model and 3D model can be directly integrated into the water conservancy digital management platform, meeting the practical needs of water conservancy engineering mapping and digital management. Attached Figure Description

[0038] Figure 1 This is a flowchart of the method of the present invention;

[0039] Figure 2 This is a system module diagram of the present invention. Detailed Implementation

[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0041] Please see Figure 1 This invention provides a method for stereo mapping and 3D modeling of water conservancy using the Gaofen-7 satellite, comprising: screening stereo images on the same track using a quantitative evaluation model of topological features of water conservancy facilities; performing stereo image stitching and fusion based on linear topology of rivers, planar topology of reservoir areas, and node topology of canals; dividing the terrain into zones and establishing a correction model for the correlation between terrain slope and RPC parameters; extracting rigid topological features of water conservancy facilities to construct constraint equations to correct the original RPC parameters and perform error compensation; constructing a virtual constraint-driven image control-free three-dimensional solution framework for water conservancy geometric features; combining a multi-scale topological feature matching algorithm for adjustment iterative solution and accuracy self-calibration; establishing a topological weight assignment system for water conservancy facilities; guiding the sparse deployment of image control points through topological weights and performing joint adjustment optimization; extracting 3D point clouds from stereo images and performing adaptive filtering; performing digital elevation model revision for feature areas to enhance elevation accuracy; performing 3D modeling and cross-scene topological alignment based on water conservancy semantic segmentation; and performing full-element topological verification and semantic information embedding on the model.

[0042] In this embodiment, the topological features of water conservancy facilities are used as the core driving force, running through the entire process of image preprocessing, aerial triangulation, image control point deployment, point cloud processing, and 3D modeling. By replacing the precision control function of traditional dense image control points with topological constraints, the problem of insufficient precision of existing satellite stereo mapping in complex water conservancy terrain, high cost of image control point deployment, and low efficiency of large-scene modeling is solved. This realizes the efficient and high-precision application of Gaofen-7 satellite imagery in water conservancy engineering mapping.

[0043] In a specific embodiment, the screening of co-track stereo images using a quantitative evaluation model of water conservancy facility topology features includes: setting a standard threshold for the overlap of stereo models, limiting the cloud coverage area of ​​single-scene images, comprehensively evaluating image contrast, noise level, and imaging angle indicators to complete the initial screening of images, and extracting regional segments with complete water conservancy topology features from the initially screened images; the co-track stereo image stitching and fusion includes completing cloudless region stitching based on water conservancy topology feature matching, generating a fused image with complete water conservancy topology through grayscale normalization processing, and establishing a feature association index between supporting parameters and images.

[0044] Specifically, during the image selection stage, priority is given to using stereo relative images from the same orbit as the Gaofen-7 satellite. The standard threshold for the overlap of the stereo model is set according to the continuity requirements of linear water conservancy facilities, for example, it is set to no less than 4%, to ensure that linear and blocky water conservancy facilities such as rivers and dams are fully represented in the stereo model. The cloud coverage area limit for a single image is determined based on the unobstructed nature of key water conservancy areas, for example, it is set to no more than 15% of the total image area, to avoid cloud shadows obscuring core surveying areas such as dam foundations and canal nodes. Image contrast, noise level, and imaging angle indicators are quantitatively evaluated through image grayscale histogram analysis and edge detection algorithms to select images with rich information, moderate contrast, and no obvious noise spots or bright areas. During the stereo image stitching and fusion process, the river centerline, reservoir boundary, and canal nodes are used as matching reference points. A feature-based image stitching algorithm is used to complete the seamless stitching of cloud-free areas of different stereo image pairs. After stitching, the entire image is subjected to grayscale normalization processing to eliminate illumination differences between different stereo image pairs. When establishing a feature association index between supporting parameters and images, satellite parameter files such as RPB, IMD, and XML are mapped one-to-one with the corresponding water conservancy feature areas of the images to ensure that subsequent parameter corrections can accurately locate the corresponding water conservancy scene.

[0045] In a specific embodiment, the terrain partitioning includes: identifying terrain partitions such as high-altitude water conveyance, hilly river channels, and plain irrigation areas; the correction of the original RPC parameters includes: extracting the river centerline, dam outline, and canal nodes as rigid topological features of water conservancy, constructing a correlation correction model between the terrain slope of each partition and the RPC parameters, performing regional iterative correction of the original RPC parameters, eliminating satellite imaging system bias through geometric constraint adjustment iteration, and achieving adaptation of RPC parameters to the topological features of the water conservancy scene.

[0046] Specifically, terrain zoning is achieved using a slope classification method based on a digital elevation model. The operational area is divided into three categories based on terrain slope values: high-altitude water conveyance, hilly river channels, and plain irrigation areas. High-altitude water conveyance areas correspond to terrain with slopes greater than 25°, hilly river channels correspond to terrain with slopes between 5° and 25°, and plain irrigation areas correspond to terrain with slopes less than 5°. The extraction of rigid topological features for hydraulic engineering employs an algorithm combining edge detection and semantic segmentation to automatically identify the precise coordinates of river centerlines, dam outlines, and canal system nodes. These features are spatially stable and do not easily change over time, making them suitable as benchmarks for parameter correction.

[0047] The terrain slope and RPC parameter correlation correction model originates from the basic theory of satellite photogrammetry. The original RPC parameters are derived based on the assumption of flat terrain, which will produce systematic errors in areas with large terrain undulations. This embodiment corrects the original RPC parameters based on the linear correlation between terrain slope and RPC parameter errors. The derivation process is as follows: Let the prediction error of RPC parameters under flat terrain be... For every 1° increase in terrain slope, the prediction error of RPC parameters increases. Then any terrain slope The RPC parameter prediction error is as follows: ,in This is the terrain slope correction factor. Therefore, the corrected RPC parameters are the original RPC parameters minus the prediction error, i.e.: .

[0048] The calculation example of this formula is as follows: Given that the average topographic slope of a hilly river channel area is S=15∘, the predicted elevation value of the original RPC parameters for this area is... The terrain slope correction coefficient k = 0.08 m / ∘ for this terrain zone was obtained by calibrating data from three stable canal system nodes through actual measurements. The prediction error under flat terrain is... Substituting into the formula, the prediction error is calculated. The corrected RPC parameter elevation prediction value is The corrected predicted value was compared with the measured value of 119.3m, and the error was 0.2m, which meets the accuracy requirements of water conservancy surveying.

[0049] When performing regional iterative correction on the original RPC parameters, the difference between the theoretical coordinates and the image coordinates of the hydraulic rigid topology features is used as the termination condition for the iteration. The iteration stops when the difference is less than a preset threshold. The satellite imaging system bias is gradually eliminated through multiple iterations.

[0050] In one specific embodiment, the image-controlled aerial triangulation framework introduces terrain constraints, hydraulic topology constraints, and publicly available terrain data constraints; the multi-scale topological feature matching algorithm performs feature fusion for water surfaces, slopes, and tidal flats to improve the stability and number of matching points; the adjustment iterative solution uses hydraulic geometric virtual feature points as constraints in the calculation, eliminating the need for ground control points to complete aerial triangulation.

[0051] Specifically, the imageless aerial tri-solution framework employs a multi-source constraint fusion adjustment strategy. Topographic constraints originate from preprocessed digital elevation model (DEM) data, while hydraulic topological constraints are derived from extracted rigid topological features such as river centerlines and dam outlines. Publicly available topographic data constraints utilize SRTM 30-meter resolution DEM data. The multi-scale topological feature matching algorithm constructs image pyramids at different scales, extracting feature points at coarse, meso, and fine scales. For water surface areas, a texture-based feature extraction method is used, while for slope and tidal flat areas, an edge-based feature extraction method is employed. Finally, feature points from different regions and scales are fused to generate a stable set of matching points.

[0052] The adjustment iteration is based on least squares adjustment theory. In this embodiment, virtual feature points of hydraulic geometry are added as additional observations to the traditional aerial triangulation adjustment equation. The improved adjustment error equation is V=AX−L, where V is the residual vector, A is the coefficient matrix, X is the image exterior orientation element vector to be determined, and L is the observation vector. The observation vector L is composed of the coordinates of the image matching point, the coordinates of the topographic constraint point, and the coordinates of the virtual feature points of hydraulic geometry. The coordinates of the virtual feature points of hydraulic geometry are derived from the hydraulic topology. For example, the coordinates of the virtual feature point of the river centerline are obtained by linear interpolation of the coordinates of two adjacent measured canal system nodes. The adjustment iteration adopts the Gauss-Newton method. Through multiple iterations, the image exterior orientation elements are adjusted to minimize the sum of squared residuals of all observations, ultimately achieving aerial triangulation without the need for ground control points.

[0053] In one specific embodiment, the accuracy self-calibration includes: collecting measured data of water conservancy points, comparing the measured data with the calculation results point by point, constructing an error feedback mechanism to adjust the calculation parameters, and forming a water conservancy imageless control map accuracy grading judgment standard.

[0054] Specifically, the measured data of water conservancy points were collected using GNSSRTK measurement. Priority was given to stable points on key water conservancy facilities such as dam crests, canal slopes, and river cross-sections. A suitable number of measured points were collected for each topographic zone for accuracy verification. The measured data were compared point-by-point with the aerial triangulation results, and the mean square error of the horizontal position and the mean square error of the elevation were calculated. When constructing the error feedback mechanism, the calculated error values ​​were fed back to the aerial triangulation module, automatically adjusting the convergence threshold and constraint weight parameters of the adjustment iteration, and re-performing the aerial triangulation until the accuracy met the preset requirements. The accuracy grading standard for water conservancy non-image control mapping was formulated based on the mapping accuracy requirements of different water conservancy projects, dividing the accuracy into multiple levels, each corresponding to different types of water conservancy project application scenarios, providing a unified accuracy judgment basis for subsequent water conservancy surveying projects.

[0055] In a specific embodiment, the topological weighting system includes: assigning high weight values ​​to dam nodes, canal intersections, and river inflection points; the sparse deployment of image control points includes: calculating the deployment density of image control points in different weight regions by combining the imaging resolution of Gaofen-7 satellite and stereo baseline parameters, and using hydraulic geometric continuity to achieve full-domain accuracy transfer; the joint adjustment optimization includes: jointly adjusting the virtual feature points of hydraulic geometry with the real image control points, and introducing accuracy loss coefficient and operation efficiency coefficient evaluation indicators to determine the sparse deployment scheme.

[0056] Specifically, the topology weighting system for water conservancy facilities assigns values ​​based on the importance and topological stability of each facility. This embodiment uses the analytic hierarchy process (AHP) to determine the weight values ​​of each facility. The derivation process is as follows: First, an importance judgment matrix for water conservancy facilities is constructed. Experts in water conservancy surveying are invited to score the relative importance of dam nodes, canal intersections, river bends, and general areas. Then, the maximum eigenvalue and corresponding eigenvector of the judgment matrix are calculated. Finally, the eigenvectors are normalized to obtain the weight values ​​of each water conservancy facility. The weight value calculation formula is as follows: ,in For the first Weight values ​​for water conservancy facilities To determine the eigenvector of a matrix One element, This refers to the number of different types of water conservancy facilities.

[0057] The calculation example of this formula is as follows: Construct a 4th-order judgment matrix, and obtain the eigenvectors through expert scoring. Substituting into the formula, the weights of the dam nodes are calculated. Weight of intersection points of canal system River inflection point weight General regional weights .

[0058] When calculating the density of image control points (ARPCs), the panchromatic imaging resolution of 0.65 meters and the stereo baseline height ratio of approximately 1.0 of the Gaofen-7 satellite are considered comprehensively. Higher density is used in high-weight areas, and lower density in low-weight areas. Utilizing the continuity of hydraulic geometry, high-precision ARPCs in high-weight areas transfer accuracy to surrounding low-weight areas, achieving a uniform distribution of accuracy across the entire region. During the joint adjustment and optimization process, virtual feature points of hydraulic geometry and real ARPCs are used as common observations in the adjustment calculation. Two evaluation indicators are introduced: an accuracy loss coefficient to measure the loss of mapping accuracy under different deployment schemes, and an operational efficiency coefficient to measure the amount of fieldwork under different deployment schemes. By comprehensively weighing these two indicators, the optimal sparse ARPC deployment scheme is determined.

[0059] In a specific embodiment, the adaptive filtering process includes: distinguishing between ground feature points and ground points based on hydraulic topological features, automatically eliminating non-topographic interference points and retaining hydraulic topographic feature points, and adjusting the point cloud thinning density according to different topographic zones; the digital elevation model revision includes: performing geometrically guided elevation revision on river cross-sections, dam foundations, and canal slopes to make the digital elevation model meet the geometric requirements of hydraulic surveying and mapping projects.

[0060] Specifically, adaptive filtering is based on the spatial geometric and hydraulic topological features of point clouds. First, it distinguishes between ground features and surface points based on the abrupt elevation changes in the point cloud. Then, combining hydraulic topological features, point clouds located near the river centerline, dam outline, and canal nodes are marked as hydraulic topographic feature points and retained, while non-topographic interference points such as vegetation and buildings are automatically removed. The point cloud thinning density is adjusted according to the degree of topographic undulation in different terrain zones. Higher point cloud densities are used in high-altitude water conveyance areas with large topographic undulations, while lower density is used in flat plain irrigation areas, reducing data volume while preserving topographic details. The digital elevation model (DEM) revision uses a geometry-guided method, guided by the extracted hydraulic topological features, to correct the elevation values ​​of key areas such as river cross-sections, dam foundations, and canal slopes point by point. The correction process strictly adheres to the geometric connections of hydraulic facilities to ensure that the corrected DEM accurately reflects the true form of the hydraulic topography and meets the requirements of the water conservancy surveying project.

[0061] In one specific embodiment, the 3D modeling includes: identifying water conservancy facilities based on water conservancy semantic segmentation and constructing semantic association rules, and fusing digital elevation models with orthophotos to generate a full-domain 3D model; the cross-scene topology alignment includes: using a water conservancy topology centerline alignment algorithm to process the stitching area of ​​multi-scene image models, eliminating stitching gaps and accuracy deviations; the full-element topology verification includes: verifying the compliance of spatial relationships, geometric dimensions, and geometric connections of water conservancy facilities.

[0062] Specifically, the water conservancy semantic segmentation is achieved using a deep learning semantic segmentation model. The training dataset contains a large number of Gaofen-7 satellite images labeled with water conservancy facilities such as dams, sluices, canals, and rivers. The model can automatically identify and classify various water conservancy facilities in the images. When constructing semantic association rules, the spatial connection relationships between different types of water conservancy facilities are defined, such as the connection relationship between canals and rivers, and the connection relationship between dams and reservoirs. The digital elevation model and orthophotos are fused through texture mapping to generate a full-domain 3D model with realistic textures. During cross-scene topology alignment, the center lines of water conservancy topology such as the center lines of rivers and canals in the overlapping area of ​​adjacent images are used as alignment benchmarks. The iterative nearest-point algorithm is used to register the two models to eliminate stitching gaps and accuracy deviations. Full-element topology verification checks the spatial positional relationships, geometric dimensional accuracy, and connection correctness of water conservancy facilities in turn. Any parts that do not meet the requirements are manually corrected to ensure the geometric consistency of the 3D model.

[0063] In one specific embodiment, semantic information embedding includes: writing water conservancy semantic information and engineering attribute data into the model to establish a mapping relationship between geometric elements and water conservancy business data; it also includes performing lightweight geometric optimization on the model and adjusting the model data format and accuracy indicators to adapt to the access standards of the water conservancy digital management platform.

[0064] Specifically, during the semantic information embedding process, engineering attribute data such as the type, number, construction time, and design parameters of water conservancy facilities are associated with corresponding geometric elements in the 3D model, establishing a one-to-one mapping relationship between geometric elements and water conservancy business data. Lightweight geometric optimization of the model is achieved by simplifying the model's geometric structure, merging duplicate vertices, and optimizing texture mapping, significantly reducing the model's data volume without affecting its visual effects and accuracy. The model's data format is adjusted to a common format supported by the water conservancy digital management platform. Simultaneously, based on the platform's accuracy requirements, the model's geometric accuracy and texture resolution are adjusted to ensure that the model can be directly connected to the water conservancy digital management platform for use in virtual simulation, operation and maintenance management, and emergency command of water conservancy projects.

[0065] Please see Figure 2This invention also provides a Gaofen-7 satellite-based water conservancy stereo mapping and 3D modeling system, comprising: a topology feature preprocessing module for water conservancy topology evaluation, image fusion, coordinate transformation, and RPC parameter fine-tuning, outputting standardized topology mapping data; a topology constraint-free imager-based solution module for executing aerial triangulation iterative algorithms and accuracy self-calibration, generating mapping result reports; a topology weighted imager-based deployment module for calculating topology weights and generating sparse imager-based point deployment schemes, outputting field operation guidance documents; a topology filtering and modeling module for performing point cloud feature filtering, model revision, semantic modeling, and cross-scene alignment; and a topology verification and result output module for performing model geometric verification, semantic embedding, lightweight optimization, and format conversion, outputting final results conforming to the water conservancy digital platform.

[0066] Specifically, the modules communicate via a gigabit Ethernet interface and use the TCP / IP protocol for data transmission. The topology feature preprocessing module transmits the filtered and preprocessed image data, the refined RPC parameter data, and the extracted hydraulic topology feature data to the topology-constrained image control-free solution module. The topology-constrained image control-free solution module transmits the aerial triangulation results data and accuracy report data to the topology-weighted image control deployment module and the topology filtering and modeling module. The topology-weighted image control deployment module transmits the generated sparse image control point deployment scheme data to the topology-constrained image control-free solution module for joint adjustment optimization. The topology filtering and modeling module transmits the initially generated 3D model data to the topology verification and results output module. The system also includes two auxiliary modules: a data storage module and a human-computer interaction module. The data storage module is connected to each core module via a SATA 3.0 interface and is used to store raw image data, intermediate processed data, and final results data. The human-computer interaction module is connected to the system host via an HDMI 2.1 interface and is used to receive user input commands and display processing progress and results.

[0067] In summary, this invention utilizes the topological features of water conservancy facilities throughout the entire process of water conservancy stereo mapping and 3D modeling. It replaces traditional dense image control points with topological constraints to achieve accuracy control, completing the accuracy verification of image control-free mapping of Gaofen-7 satellite imagery in complex water conservancy terrains. By combining satellite imaging characteristics with the distribution patterns of water conservancy facilities, the invention optimizes the layout of image control points, significantly reducing the cost and operational risks of image control point deployment in remote areas. Simultaneously, it improves the modeling coverage and processing efficiency of large-scale water conservancy scenarios such as large reservoirs and long-distance water conveyance lines, fully leveraging the mapping capabilities of Gaofen-7 satellite in water conservancy projects. This effectively improves the accuracy and operational efficiency of water conservancy engineering mapping and modeling, and also establishes an accuracy judgment standard for image control-free mapping adapted to water conservancy scenarios. The optimized digital elevation model and 3D model can be directly integrated into the water conservancy digital management platform, meeting the practical needs of water conservancy engineering mapping and digital management.

[0068] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0069] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for three-dimensional mapping and modeling of water conservancy systems using the Gaofen-7 satellite, characterized in that, include: A quantitative evaluation model of the topological characteristics of water conservancy facilities was used to screen the same-track stereo images, and the same-track stereo images were stitched and fused based on the linear topology of the river channel, the planar topology of the reservoir area, and the node topology of the canal system. The terrain is divided into zones and a correction model is established to correlate terrain slope with RPC parameters. The rigid topological features of hydraulic engineering are extracted to construct constraint equations to correct the original RPC parameters and perform error compensation. A virtual constraint-driven, imageless, three-solution framework for hydraulic geometric features is constructed, and a multi-scale topological feature matching algorithm is used for adjustment iterative solution and accuracy self-calibration. Establish a topology weighting system for water conservancy facilities, and guide the sparse deployment of image control points and perform joint adjustment optimization through topology weights; Extract the 3D point cloud from the stereo image and perform adaptive filtering. Then, perform digital elevation model revision for the feature regions to enhance elevation accuracy. Based on water conservancy semantic segmentation, 3D modeling and cross-scene topology alignment are performed, and full-element topology verification and semantic information embedding are performed on the model.

2. The method for three-dimensional mapping and modeling of water conservancy systems using the Gaofen-7 satellite according to claim 1, characterized in that, The method of using the quantitative evaluation model of topological features of water conservancy facilities to screen the same-track stereo images includes: setting a standard threshold for the overlap of stereo models, limiting the cloud coverage area of ​​a single scene image, comprehensively evaluating the image contrast, noise level and imaging angle indicators to complete the initial screening of images, and extracting regional segments with complete water conservancy topological features from the initial screened images. The same-track stereo image stitching and fusion includes stitching cloudless areas based on hydraulic topology feature matching, generating a complete hydraulic topology fused image through grayscale normalization processing, and establishing a feature association index between supporting parameters and images.

3. The method for three-dimensional mapping and modeling of water conservancy systems using the Gaofen-7 satellite according to claim 1, characterized in that, The division of terrain zones includes: identifying terrain zones such as high-mountain water conveyance, hilly river channels, and plain irrigation areas; The correction of the original RPC parameters includes: extracting the river centerline, dam outline, and canal nodes as rigid topological features of the water conservancy system; constructing a correlation correction model between the terrain slope of each region and the RPC parameters; performing regional iterative correction of the original RPC parameters; eliminating satellite imaging system bias through geometric constraint adjustment iteration; and achieving the adaptation of RPC parameters to the topological features of the water conservancy scene.

4. The method for three-dimensional mapping and modeling of water conservancy systems using the Gaofen-7 satellite according to claim 1, characterized in that, The imageless spatial three-solution framework introduces terrain constraints, hydraulic topology constraints, and publicly available terrain data constraints. The multi-scale topological feature matching algorithm performs feature fusion on water surface, slope and tidal flat areas to improve the stability and number of matching points. The adjustment iterative solution uses virtual features of hydraulic topology as constraints in the calculation, eliminating the need for ground control points to complete aerial triangulation.

5. The method for three-dimensional mapping and modeling of water conservancy systems using the Gaofen-7 satellite according to claim 1, characterized in that, The accuracy self-calibration includes: collecting measured data of water conservancy points, comparing the measured data with the calculation results point by point, constructing an error feedback mechanism to adjust the calculation parameters, and forming a standard for judging the accuracy of water conservancy imageless control mapping.

6. The method for three-dimensional mapping and modeling of water conservancy systems using the Gaofen-7 satellite according to claim 1, characterized in that, The topological weighting system includes assigning high weight values ​​to dam nodes, canal intersections, and river bends. The sparse deployment of image control points includes: calculating the deployment density of image control points in different weighted regions by combining the imaging resolution of Gaofen-7 satellite and stereo baseline parameters, and using hydraulic geometric continuity to achieve full-domain accuracy transfer; The adjustment and joint optimization includes: jointly adjusting the virtual feature points of hydraulic geometry with the real control points, and introducing the accuracy loss coefficient and operation efficiency coefficient as evaluation indicators to determine the sparse layout scheme.

7. The method for three-dimensional mapping and modeling of water conservancy systems using the Gaofen-7 satellite according to claim 1, characterized in that, The adaptive filtering process includes: distinguishing between ground feature points and ground points based on hydraulic topology features, automatically eliminating non-topographic interference points and retaining hydraulic topographic feature points, and adjusting the point cloud thinning density according to different topographic zones. The digital elevation model revision includes: performing geometrically guided elevation revision on river cross sections, dam foundations, and canal slopes to ensure that the digital elevation model meets the geometric requirements of water conservancy surveying and mapping.

8. The method for three-dimensional mapping and modeling of water conservancy systems using the Gaofen-7 satellite according to claim 1, characterized in that, The three-dimensional modeling includes: identifying water conservancy facilities based on water conservancy semantic segmentation and constructing semantic association rules, and fusing digital elevation models with orthophotos to generate a full-domain three-dimensional model; The cross-scene topology alignment includes: using a hydraulic topology centerline alignment algorithm to process the stitching area of ​​multi-scene image models, eliminating stitching gaps and accuracy deviations; The full-element topology verification includes: checking the compliance of the spatial relationships, geometric dimensions, and geometric connections of water conservancy facilities.

9. A method for three-dimensional mapping and modeling of water conservancy systems using the Gaofen-7 satellite according to claim 1, characterized in that, The semantic information embedding includes: writing water conservancy semantic information and engineering attribute data into the model, and establishing a mapping relationship between geometric elements and water conservancy business data; It also includes lightweight geometric optimization of the model and adjustment of the model data format and accuracy indicators to adapt to the access standards of the water conservancy digital management platform.

10. A Gaofen-7 satellite water conservancy stereo mapping and 3D modeling system, applied to the Gaofen-7 satellite water conservancy stereo mapping and 3D modeling method according to any one of claims 1 to 9, characterized in that, include: The topology feature preprocessing module is used for hydraulic topology evaluation, image fusion, coordinate transformation, and RPC parameter fine-tuning, and outputs standardized topology mapping data. The topology-constrained image control-free solution module is used to execute the aerial triangulation iterative algorithm and accuracy self-calibration, and generate mapping results reports; The topology weighting and sparse control point deployment module is used to calculate topology weights and generate sparse control point deployment schemes, and output field operation guidance documents. The topology filtering and modeling module is used to perform point cloud topology filtering, model revision, semantic modeling, and cross-scene alignment. The topology verification and output module is used to perform model topology verification, semantic embedding, lightweight optimization, and format conversion, and output the final results that conform to the water conservancy digital platform.

Images