Point cloud data processing method and device of laser radar, equipment and medium
By using multi-dimensional geometric feature collaborative analysis and dynamic weight optimization, the limitations of traditional point cloud denoising methods in complex scenarios are solved, achieving efficient denoising of point cloud data and accurate preservation of key structures.
Patent Information
- Application Number
- CN202511937046.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-22
- Publication Date
- 2026-01-16
- Estimated Expiration
- 2045-12-22
AI Technical Summary
Traditional point cloud denoising methods are difficult to adapt to changes in local features in regions of different densities, resulting in the accidental deletion of valid points in sparse regions and the omission of noise in dense regions. Furthermore, they lack a comprehensive understanding of the local morphology of point clouds, leading to poor preservation of fine structures. Parameter settings rely on human experience, resulting in poor versatility and robustness.
By employing collaborative analysis of multi-dimensional geometric features and a dynamic weight optimization mechanism, and using adaptive local neighborhood construction, covariance matrix decomposition, dynamic weight optimization, and clustering algorithms, intelligent identification and denoising of point clouds and accurate preservation of effective point clouds are achieved.
It improves the effectiveness of feature fusion and anti-interference capability in point cloud data processing, and achieves efficient denoising and accurate preservation of key structures in complex real-world scenarios.