Local map point cloud global alignment method and device, electronic equipment and storage medium

By constructing a cross-modal, multi-scale hypergraph structure to align local map point clouds and remote sensing images, the problems of insufficient positioning accuracy and high economic cost were solved, and high-precision point cloud positioning was achieved.

CN117351047BActive Publication Date: 2026-06-26TSINGHUA UNIVERSITY
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
TSINGHUA UNIVERSITY
Filing Date
2023-08-22
Publication Date
2026-06-26

AI Technical Summary

Technical Problem

The existing image-based positioning method increases the conversion error from image to point cloud data, resulting in insufficient positioning accuracy. Furthermore, using panoramic cameras for data acquisition increases economic costs.

Method used

By acquiring local ground point clouds and remote sensing images of the target area to be registered, geometric transformation and feature extraction are performed to construct a cross-modal multi-scale hypergraph structure. An orientation mask is used for fitting and matching, the best matching relationship is calculated, and the point cloud and remote sensing image are aligned to the same geometric space.

Benefits of technology

It improves positioning accuracy, reduces economic costs, and solves the problem of image-to-point cloud data conversion error.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN117351047B_ABST
    Figure CN117351047B_ABST
Patent Text Reader

Abstract

The application relates to the field of computer vision, in particular to a local map point cloud global alignment method and device, electronic equipment and a storage medium. The method comprises the following steps: performing feature extraction and representation on target area to-be-registered ground local point cloud data and a global remote sensing image, and constructing a cross-modal feature matching model, and obtaining the relative position and direction of the ground local point cloud in the global remote sensing image through matching, and aligning the map data of two modes and different scales to the same geometric space. Thus, through semantic representation of the local ground point cloud data and the large-scale remote sensing image, the relative position and direction of the local point cloud in the global remote sensing image are obtained, the conversion error from the image to the point cloud data is increased in the current image positioning mode, the positioning accuracy is insufficient, and the panoramic image needs to be collected by a special panoramic camera, thereby increasing the economic cost and other problems. The positioning accuracy is improved, and the economic cost is reduced.
Need to check novelty before this filing date? Find Prior Art