A point cloud registration driven underwater structure deformation measurement method
By using an initial transformation-guided and Gaussian probability density function-driven fine registration method, combined with Gaussian mixture model and K-nearest neighbor method, the problem of low accuracy of traditional point cloud registration methods for underwater strip point clouds is solved, and full-range, accurate underwater deformation monitoring is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- XIAN INST OF OPTICS & PRECISION MECHANICS CHINESE ACAD OF SCI
- Filing Date
- 2026-03-30
- Publication Date
- 2026-07-17
AI Technical Summary
Traditional point cloud registration methods are not optimized for the slender geometric characteristics of strip point clouds, resulting in low accuracy in underwater strip structure deformation measurement, which cannot meet the needs of precise monitoring.
A method guided by initial transformation and driven by Gaussian probability density function for precise registration is adopted. Point cloud data is registered using Gaussian mixture model and K-nearest neighbor method. Combined with laser triangulation measurement system, accurate matching and deformation measurement of strip point clouds are achieved.
It significantly improves the accuracy of underwater strip structure deformation measurement, realizes full-range and accurate deformation monitoring, and provides reliable data support for the safe operation and maintenance of underwater equipment.
Smart Images
Figure CN122415690A_ABST