Search

CN-122023693-A - Reconstruction method, device, equipment, medium and product of three-dimensional terrain model

CN122023693ACN 122023693 ACN122023693 ACN 122023693ACN-122023693-A

Abstract

The disclosure relates to the technical field of three-dimensional terrain modeling, and provides a reconstruction method, a device, equipment, a medium and a product of a three-dimensional terrain model. The three-dimensional terrain model reconstruction method comprises the steps of firstly obtaining a height image representing the height of a target three-dimensional terrain model, then creating a corresponding target point cloud based on the height image, enabling the height value of a space point in the point cloud to correspond to the height of the terrain, further, carrying out updating operation on the space point in the target point cloud, wherein the updating operation is completed by utilizing the height value of other space points in the neighborhood range of each space point, and finally, reconstructing the target three-dimensional terrain model by using the updated target point cloud. The present disclosure addresses the problem of surface imperfections and non-smoothness during three-dimensional terrain reconstruction in the related art.

Inventors

  • SHEN ZHONGYONG

Assignees

  • 广州博冠信息科技有限公司

Dates

Publication Date
20260512
Application Date
20260109

Claims (13)

  1. 1. A method for reconstructing a three-dimensional terrain model, the method comprising: generating a height image corresponding to a target three-dimensional terrain model, wherein the height image is used for representing the heights of a plurality of position points on the target three-dimensional terrain model; Creating a target point cloud corresponding to the height image, wherein the height value of a space point in the target point cloud corresponds to the height of the position point; updating the height value of each space point in the target point cloud by utilizing the height values of a plurality of space points in the neighborhood range of the space point; and reconstructing the target three-dimensional terrain model based on the updated target point cloud.
  2. 2. The method of claim 1, wherein generating the height image corresponding to the target three-dimensional terrain model comprises: extracting the height information of a plurality of position points on the target three-dimensional terrain model; And carrying out normalization processing on the height information of the plurality of position points, and encoding the normalized height information into a height image, wherein the height image is an EXR format image with floating point precision, and the height information is stored in a target color channel of the EXR format image.
  3. 3. The method of claim 2, wherein extracting altitude information for a plurality of location points on the target three-dimensional terrain model comprises: Scanning the target three-dimensional terrain model through an orthogonal camera in a first three-dimensional platform to obtain the world space position of each position point on the target three-dimensional terrain model under the view angle of the orthogonal camera; and extracting the height value of the world space position in the preset axial direction to obtain the height information.
  4. 4. The method of claim 1, wherein the creating the target point cloud for the height image comprises: generating an initial point cloud in a second three-dimensional platform according to the height information in the height image; Generating a corresponding geometric patch for each point in the initial point cloud; The target point cloud is generated based on a plurality of vertices of the geometric patch.
  5. 5. The method of claim 4, wherein generating an initial point cloud from the height information in the height image comprises: the height image is converted to a height field volume, and the height field volume is converted to points, thereby generating the initial point cloud.
  6. 6. The method of claim 4 or 5, wherein the generating the target point cloud based on the plurality of vertices of the geometric patch comprises: adjacent vertices of a plurality of geometric patches are stitched to construct the cloud of target points having a topological neighborhood relationship.
  7. 7. The method of claim 4, wherein the updating with the height values of the plurality of spatial points within the neighborhood of the spatial point for each spatial point in the target point cloud comprises: For each current space point in the target point cloud, acquiring all adjacent space points of the current space point in a preset neighborhood range; Obtaining an average height value based on the height values of all the adjacent space points; And updating the height value of the current space point based on the average height value.
  8. 8. The method of claim 7, wherein the deriving an average height value based on the height values of all the neighboring spatial points comprises: Performing inverse distance weighting calculation on the height values of all the adjacent space points to obtain weight data of each adjacent space point in all the adjacent space points; And obtaining the average height value according to the height value and the weight data of each adjacent space point in all the adjacent space points.
  9. 9. The method according to claim 7 or 8, wherein reconstructing the target three-dimensional terrain model based on the updated target point cloud comprises: Transmitting the updated height value of the current space point back to the corresponding vertex in the geometric patch in an attribute transmission mode so as to update the height value of the geometric patch; And reconstructing to obtain the target three-dimensional terrain model based on the geometrical patch with the updated height value.
  10. 10. A reconstruction apparatus for a three-dimensional terrain model, the apparatus comprising: The generation module is used for generating a height image corresponding to the target three-dimensional terrain model, wherein the height image is used for representing the heights of a plurality of position points on the target three-dimensional terrain model; The creation module is used for creating a target point cloud corresponding to the height image, and the height value of a space point in the target point cloud corresponds to the height of the position point; the updating module is used for updating the height value of each space point in the target point cloud by utilizing the height values of a plurality of space points in the neighborhood range of the space point; and the reconstruction module is used for reconstructing the target three-dimensional terrain model based on the updated target point cloud.
  11. 11. An electronic device, comprising: A memory and a processor, the memory and the processor being communicatively connected to each other, the memory having stored therein computer instructions, the processor executing the computer instructions to perform the method of reconstructing a three-dimensional terrain model as claimed in any of claims 1 to 9.
  12. 12. A computer readable storage medium having stored thereon computer instructions for causing a computer to perform the method of reconstructing a three-dimensional terrain model according to any of claims 1 to 9.
  13. 13. A computer program product comprising computer instructions for causing a computer to perform the method of reconstructing a three-dimensional terrain model according to any of claims 1 to 9.

Description

Reconstruction method, device, equipment, medium and product of three-dimensional terrain model Technical Field The disclosure relates to the technical field of three-dimensional terrain modeling, in particular to a method, a device, equipment, a medium and a product for reconstructing a three-dimensional terrain model. Background In the existing three-dimensional terrain modeling, a common method is to extract the height information of the terrain, generate a corresponding height image and reconstruct a three-dimensional model according to the information in the height image. However, this method of generating a three-dimensional terrain representation directly from a height image often results in a reconstructed three-dimensional terrain representation that is not smooth enough in surface due to the discretization of the two-dimensional image pixels, with visual imperfections that affect the quality and realism of the final reconstructed model. Therefore, a method for reconstructing a three-dimensional terrain model is needed to solve the problems of surface flaws and non-smoothness during three-dimensional terrain reconstruction in the related art. Disclosure of Invention The disclosure provides a reconstruction method, a device, equipment, a medium and a product of a three-dimensional terrain model, which are used for solving the problems of surface flaws and unsmooth in the reconstruction of the three-dimensional terrain in the related technology. In a first aspect, the present disclosure provides a method of reconstructing a three-dimensional terrain model, the method comprising. Generating a height image corresponding to the target three-dimensional terrain model, wherein the height image is used for representing the heights of a plurality of position points on the target three-dimensional terrain model; creating a target point cloud corresponding to the height image, wherein the height value of a space point in the target point cloud corresponds to the height of a position point; Updating the height value of each space point in the target point cloud by utilizing the height values of a plurality of space points in the neighborhood range of the space point; And reconstructing the target three-dimensional terrain model based on the updated target point cloud. According to the reconstruction method of the three-dimensional terrain model, after the target point cloud is created based on the height image, the target point cloud is not directly utilized for reconstruction, the height value of each space point in the target point cloud is updated by utilizing the height values of a plurality of space points in the neighborhood range of the space point, and the updating step is essentially that the processing and optimization of the point cloud data based on neighborhood information are carried out once. This process can smooth or correct the original spatial point values that may have noise or abrupt changes that are directly converted from the height image. Therefore, the method and the device reconstruct based on the updated target point cloud with higher quality, can effectively improve the surface smoothness and the data accuracy of the finally generated three-dimensional terrain model, and reduce the defects possibly generated by direct reconstruction. In an alternative embodiment, generating a height image corresponding to the target three-dimensional terrain model includes: extracting the height information of a plurality of position points on the target three-dimensional terrain model; and carrying out normalization processing on the height information of the plurality of position points, and encoding the normalized height information into a height image, wherein the height image is an EXR format image with floating point precision, and the height information is stored in a target color channel of the EXR format image. By adopting the floating point EXR format supporting high dynamic range and high precision to store the height data, the method can keep details and precision in the original height information of the terrain to the maximum extent, avoid the loss of the height information caused by data compression or format limitation in the generation stage of the height image, and provide a high-fidelity data base for the subsequent point cloud reconstruction. In an alternative embodiment, extracting altitude information for a plurality of location points on a target three-dimensional terrain model includes: scanning the target three-dimensional terrain model through an orthogonal camera in a first three-dimensional platform to obtain the world space position of each position point on the target three-dimensional terrain model under the view angle of the orthogonal camera; and extracting the height value of the world space position in the preset axial direction to obtain the height information. The method and the device also utilize the characteristic that the orthogonal camera has no perspective