A method and system for extracting a texture image based on a three-dimensional point cloud

By extracting the basic planar point cloud set from 3D point cloud data and calculating the distance to generate the planar point cloud set, using the neighborhood point density to identify the boundary point cloud, determining the order of the circumscribed rectangle and ring, establishing a 2D coordinate system, obtaining the projected coordinates and generating the texture image, the problem of poor robustness and limited adaptability of texture image extraction in the prior art is solved, and efficient and accurate texture image extraction is achieved.

CN122289727APending Publication Date: 2026-06-26GUANGZHOU POWER SUPPLY BUREAU GUANGDONG POWER GRID CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
GUANGZHOU POWER SUPPLY BUREAU GUANGDONG POWER GRID CO LTD
Filing Date
2026-03-26
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

Existing texture image extraction techniques are not robust in 3D point clouds and are easily affected by noise, occlusion and density unevenness, resulting in texture distortion and breakage. They are computationally complex, have limited adaptability, and require additional reliance on high-resolution textures and manual intervention, making it difficult to meet practical engineering needs.

Method used

By extracting the basic planar point cloud set from the 3D point cloud dataset, calculating the point cloud distance to generate the planar point cloud set, using the neighborhood point density to identify the boundary point cloud, determining the order of the circumscribed rectangle and ring, establishing a 2D coordinate system, obtaining the projected coordinates and generating the texture image, reducing noise interference, and improving the integrity and continuity of the texture region.

Benefits of technology

It enhances the accuracy and efficiency of texture image extraction, reduces reliance on high-resolution textures, improves adaptability to complex scenes, reduces computational complexity, reduces human intervention, and increases processing speed.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122289727A_ABST
    Figure CN122289727A_ABST
Patent Text Reader

Abstract

This invention discloses a method and system for texture image extraction based on 3D point clouds. The method includes clustering and dividing an acquired 3D point cloud dataset to generate several facade point cloud sets; obtaining the neighborhood point density of each facade point cloud in the facade point cloud sets to determine several boundary point clouds corresponding to the facade point cloud sets; determining the bounding rectangle corresponding to each facade point cloud set based on the boundary point clouds to obtain the projected coordinates of each boundary point cloud, and obtaining the two-dimensional coordinate system of each facade point cloud set based on the projected coordinates; obtaining the two-dimensional facade coordinates corresponding to each facade point cloud based on the two-dimensional coordinate system, and combining it with a preset projection resolution to obtain the pixel coordinates corresponding to each facade point cloud; obtaining the color corresponding to each facade point cloud from the 3D point cloud dataset, and generating a texture image corresponding to each facade point cloud set based on the pixel coordinates and color, thereby improving the accuracy and efficiency of texture image extraction.
Need to check novelty before this filing date? Find Prior Art