A relative localization method for wall-climbing robots in sparse feature environments based on multi-view surround view and its application

By constructing a multi-view sparse feature environment for wall-climbing robot localization, multiple independent binocular vision front-ends and inertial measurement units are fused together, which solves the problem of unstable robot localization in sparse feature environments and realizes stable and continuous pose estimation in large and complex curved surface scenes.

CN122130085APending Publication Date: 2026-06-02HUAZHONG UNIV OF SCI & TECH
View PDF 0 Cites 0 Cited by

Patent Information

Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
HUAZHONG UNIV OF SCI & TECH
Filing Date
2026-03-17
Publication Date
2026-06-02

AI Technical Summary

Technical Problem

Existing technologies struggle to achieve stable localization of adsorption-type mobile robots in sparse feature environments, especially in large and complex curved surface scenarios. Monocular or binocular visual inertial odometry is prone to losing tracking in areas with weak texture, inertial measurement unit integral drift is severe, and existing multi-view systems lack initialization robustness.

Method used

A wall-climbing robot localization method based on sparse feature environment with multi-view surround view is proposed. By constructing multiple independent binocular vision front-ends and fusing them with inertial measurement units, the visual scale can be rapidly recovered and reliably initialized. A redundant and complementary visual observation system is constructed. Combined with visual inertial odometry and zero-speed update constraints, inertial drift is suppressed.

Benefits of technology

Stable relative positioning of the robot was achieved in a sparse feature environment, which enhanced the robustness and engineering practicality of the system, ensured continuous and reliable pose estimation of the robot on curved surfaces, and adapted to the rapid deployment requirements in complex scenarios.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN122130085A_ABST
    Figure CN122130085A_ABST
Patent Text Reader

Abstract

This invention belongs to the field of robot localization and multi-sensor fusion technology, specifically relating to a localization method for adsorption-type mobile robots in sparse feature environments suitable for large and complex curved surface operations. It employs a 360° surround-view perception system and simultaneously constructs multiple sets of binocular visual inertial odometry front-ends, obtaining accurate and stable relative positioning information for the robot without relying on external sensors. By constructing a redundant and complementary visual observation system through a multi-view surround-view system, it overcomes the problem that existing single or binocular visual inertial odometry systems struggle to achieve stable and continuous feature point acquisition and visual pose estimation in such scenarios. The localization system of this invention has a low overall cost, good compatibility with the main structure of adsorption-type mobile robots, and is suitable for applications of adsorption-type mobile robots in the manufacturing and maintenance of large curved surface components such as aircraft and wind turbine blades, providing the robot with real-time, accurate, and stable relative pose information.
Need to check novelty before this filing date? Find Prior Art