Search

CN-122015766-A - Mountain area elevation mapping method based on data fusion

CN122015766ACN 122015766 ACN122015766 ACN 122015766ACN-122015766-A

Abstract

The invention relates to the technical field of surveying and mapping engineering, and discloses a mountain area elevation surveying and mapping method based on data fusion. The method constructs a space reference network formed by a plurality of control points, and acquires multi-level topographic observation data covering the network through multi-sensor equipment. And after the data are spatially registered, extracting the topographic feature points and forming a preliminary feature set, and accordingly establishing a dynamic elevation reference plane. And (3) identifying the elevation difference region and carrying out feature clustering by comparing the registration data with the reference surface, and dividing different topographic feature clusters. And aiming at each characteristic cluster, adaptively selecting an optimal data fusion algorithm to generate a corresponding refined terrain curved surface. And (3) splicing and smoothing all the curved surfaces, and outputting high-precision mountain area elevation mapping results. According to the invention, through the self-adaptive fusion of the dynamic reference plane construction and the feature clusters, the applicability limitation of the traditional single model in complex mountain areas is overcome, and the mapping precision and reliability are obviously improved.

Inventors

  • LIU ZHENGUO

Assignees

  • 江苏建筑职业技术学院

Dates

Publication Date
20260512
Application Date
20260211

Claims (10)

  1. 1.A mountain area elevation mapping method based on data fusion, the method comprising: constructing a spatial reference network for a target mountain area, wherein the spatial reference network is composed of a plurality of control points distributed in the mountain area; acquiring multi-level topographic observation data covering the space reference network through acquisition equipment carrying different sensors; carrying out space registration on the multi-level topographic observation data to generate registered observation data with a unified coordinate frame; Extracting topographic feature points and feature lines in the registered observation data to form a preliminary topographic feature set; establishing a dynamic elevation reference plane in a space reference network based on the preliminary topography feature set; comparing the registered observation data with a dynamic elevation reference surface, and identifying an elevation difference region between the observation data and the reference surface; Carrying out feature clustering on the identified elevation difference areas, and dividing different topographic feature clusters; For each topographic feature cluster, adaptively selecting a corresponding data fusion algorithm; Executing a data fusion algorithm to generate a fine topographic curved surface corresponding to the feature cluster; and (3) splicing and smoothing all the fine terrain curved surfaces, and outputting the final mountain area elevation mapping result.
  2. 2. The mountainous area elevation mapping method based on data fusion of claim 1, wherein constructing a spatial reference network for a target mountainous area comprises: determining a control point density distribution scheme of a space reference network according to the longitude and latitude range and the terrain complexity of a target mountain area; presetting theoretical layout positions of control points on a mountain topographic map according to a control point density distribution scheme; correcting the theoretical layout position by combining the accessibility of the field survey, and determining the field position of a final control point; burying a measurement mark at the final control point site, and measuring the three-dimensional coordinates of each control point by adopting a precise measurement technology; And recording the coordinate information of all the control points and the attribute metadata thereof to form a spatial reference network database.
  3. 3. The mountain area elevation mapping method based on data fusion of claim 2, wherein the acquisition of multi-level topographic observation data through the acquisition device with different sensors comprises: a scheduling satellite remote sensing platform acquires wide-area optical images and radar elevation data; Controlling an aerial photogrammetry system to acquire high-resolution aerial images and laser point cloud data; Deploying a ground movement measurement system to obtain fine three-dimensional scanning data of a key area; Synchronously recording space-time reference parameters and sensor calibration parameters of various acquisition devices; And converting the acquired original observation data into an intermediate data product in a standard format.
  4. 4. A mountainous area elevation mapping method based on data fusion according to claim 3, characterized in that spatial registration of multi-level topographic observation data comprises: Taking control point coordinates in a spatial reference network database as absolute references; performing geometric fine correction on the wide-area optical image and radar elevation data acquired by the satellite remote sensing platform; Carrying out regional network adjustment processing on the high-resolution aerial image and the laser point cloud data acquired by the aerial photogrammetry system; carrying out coordinate system unified conversion on fine three-dimensional scanning data acquired by a ground mobile measurement system; And (3) performing accuracy verification on the processed data by using the control point coordinates, and ensuring that all the data meet the uniform coordinate frame accuracy requirement.
  5. 5. The mountain area elevation mapping method based on data fusion of claim 4, wherein extracting topographical feature points and feature lines in the post-registration observation data comprises: extracting topographic abrupt line features from the optical image by adopting an edge detection operator; identifying terrain feature points from the laser point cloud data through point cloud density analysis; extracting ridge lines and valley line features from the three-dimensional scan data based on curvature calculation; constructing topological relation of the extracted various features, and eliminating redundancy and contradictory features; And carrying out attribute assignment on the reserved features to form a preliminary topography feature set with elevation values and confidence degrees.
  6. 6. The method for mapping mountainous areas elevation based on data fusion of claim 5, wherein establishing a dynamic elevation reference surface based on a preliminary topography feature set comprises: performing spatial interpolation on the preliminary terrain feature set to generate an initial elevation curved surface model; calculating the residual error of the initial elevation curved surface model and the elevation of the control point of the space reference network; establishing an elevation correction field function according to residual error distribution; optimizing an initial elevation curved surface model by applying an elevation correction field function to generate a dynamic elevation reference surface; Discretizing the dynamic elevation datum plane into a regular grid model, and facilitating subsequent comparison and analysis.
  7. 7. The method of claim 6, wherein identifying areas of elevation differences between the observed data and the reference surface comprises: resampling the registered observation data to the same spatial resolution as the dynamic elevation reference plane; calculating the difference value between the elevation value of the registered observation data and the elevation value of the dynamic elevation reference surface by grids; setting an elevation difference threshold value, and marking a grid exceeding the threshold value as an elevation abnormal grid; Performing spatial cluster analysis on the elevation abnormal grid to form a continuous elevation difference region; the boundary range and average differential intensity of each elevation difference region are recorded.
  8. 8. The method for mapping mountainous areas elevation based on data fusion of claim 7, wherein feature clustering the identified elevation difference areas comprises: calculating the topography statistical characteristics of each elevation difference region, including elevation standard deviation, gradient distribution and roughness index; performing dimension reduction treatment on the terrain statistical features by using a principal component analysis method; performing unsupervised classification on the feature vectors subjected to dimension reduction by adopting a clustering algorithm; dividing the elevation difference area into different topographic feature clusters according to the classification result; A feature template is built for each topographical feature cluster describing its typical topographical features.
  9. 9. The mountainous area elevation mapping method based on data fusion of claim 8, the method is characterized by adaptively selecting a corresponding data fusion algorithm, and comprising the following steps: establishing a data fusion algorithm library which comprises a plurality of fusion methods applicable to different topographic features; matching the feature templates of the topographic feature clusters with algorithm application conditions in a data fusion algorithm library; distributing a proper fusion algorithm and a proper parameter combination for each topographic feature cluster according to the matching degree; for special topography feature clusters, starting a multi-algorithm voting mechanism to determine a final fusion scheme; a customized fusion process flow for each topographical feature cluster is generated.
  10. 10. The method of claim 9, wherein performing a data fusion algorithm and outputting a final result comprises: According to the customized fusion processing flow, carrying out fusion calculation on the observed data in each topographic feature cluster; Adopting a progressive encryption strategy to carry out fusion result smooth transition treatment on the boundary region of the feature cluster; Seamless splicing is carried out on the fusion result of each feature cluster according to the space position; carrying out integrity check and topology consistency verification on the spliced integral elevation model; and outputting the final mountain area elevation mapping result into a digital product in a specified format after the verification is passed.

Description

Mountain area elevation mapping method based on data fusion Technical Field The invention relates to the technical field of surveying and mapping engineering, in particular to a mountain area elevation surveying and mapping method based on data fusion. Background The current mountain area height Cheng Cehui is mainly obtained by adopting a total station, a GPS and other single sensors to collect data, and the elevation calculation is carried out through a static reference plane. The prior art has poor adaptability to complex terrains, and the precision of special terrains such as steep slopes, valleys and the like is insufficient due to the fact that the datum plane is fixed. The multi-level observation data registration method is simple, and the data fusion effect of different sensors is not ideal. Feature extraction dimensions are limited and the morphological structure of the terrain cannot be fully utilized. The cluster analysis strategy is solidified, and the fusion algorithm cannot be adaptively adjusted according to the terrain change. The existing method needs to solve the key technical problems of dynamic reference plane construction, multi-source data collaborative fusion, terrain self-adaptive processing and the like. Traditional elevation mapping methods have significant shortcomings in terms of accuracy and efficiency of complex terrain areas. The spatial reference network is sparsely distributed, and the reference surface is distorted due to uneven distribution of control points. The data acquisition of the sensors is asynchronous, and the space-time consistency is difficult to ensure. The registration algorithm has low precision and large error of coordinate system unification. The feature point line is incompletely extracted, and the detail features of the topography are seriously lost. The elevation difference region is divided into coarse areas, and the boundary blurring influences the clustering effect. The fusion algorithm is selected to be empirical, and the special terrain treatment effect is poor. The curved surface splice gap is obvious, and the smooth treatment damages the authenticity of the terrain. The output formats of the results are not uniform, and the compatibility of the subsequent applications is insufficient. Disclosure of Invention The invention aims to provide a mountain area elevation mapping method based on data fusion, which aims to solve the problems in the background technology. In order to achieve the above purpose, the present invention provides a mountain area elevation mapping method based on data fusion, the method comprising: constructing a spatial reference network for a target mountain area, wherein the spatial reference network is composed of a plurality of control points distributed in the mountain area; acquiring multi-level topographic observation data covering the space reference network through acquisition equipment carrying different sensors; carrying out space registration on the multi-level topographic observation data to generate registered observation data with a unified coordinate frame; Extracting topographic feature points and feature lines in the registered observation data to form a preliminary topographic feature set; establishing a dynamic elevation reference plane in a space reference network based on the preliminary topography feature set; comparing the registered observation data with a dynamic elevation reference surface, and identifying an elevation difference region between the observation data and the reference surface; Carrying out feature clustering on the identified elevation difference areas, and dividing different topographic feature clusters; For each topographic feature cluster, adaptively selecting a corresponding data fusion algorithm; Executing a data fusion algorithm to generate a fine topographic curved surface corresponding to the feature cluster; and (3) splicing and smoothing all the fine terrain curved surfaces, and outputting the final mountain area elevation mapping result. Preferably, constructing a spatial reference network for the target mountain area includes: determining a control point density distribution scheme of a space reference network according to the longitude and latitude range and the terrain complexity of a target mountain area; presetting theoretical layout positions of control points on a mountain topographic map according to a control point density distribution scheme; correcting the theoretical layout position by combining the accessibility of the field survey, and determining the field position of a final control point; burying a measurement mark at the final control point site, and measuring the three-dimensional coordinates of each control point by adopting a precise measurement technology; And recording the coordinate information of all the control points and the attribute metadata thereof to form a spatial reference network database. Preferably, the acquisition device for carrying differen