一种交通事故现场重建方法及装置
By constructing a keyframe set and a fusion optimization model, and combining it with the dynamic changes of airborne light sources, the problem of insensitivity to microscopic morphology in existing technologies has been solved, achieving high-precision 3D reconstruction of traffic accident scenes, capturing key details, and improving the accuracy of reconstruction.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- ZHEJIANG EXPRESSWAY CO LTD
- Filing Date
- 2026-02-26
- Publication Date
- 2026-07-17
AI Technical Summary
Existing methods for reconstructing traffic accident scenes neglect changes in lighting in videos captured by drones, resulting in insensitivity to microscopic features, loss of crucial details, and reduced accuracy of reconstruction.
By acquiring video frame sequences collected by drones, a set of keyframes is constructed. Combined with the dynamic changes of airborne light sources, macroscopic geometric reconstruction and microscopic normal recovery are performed. A fusion optimization model is constructed, and the macroscopic depth map and microscopic normal field are fused to generate a fused depth map. Finally, an accident scene reconstruction model is constructed.
It enables sub-centimeter-level 3D reconstruction without the need for external auxiliary equipment, effectively capturing key details of the accident scene, restoring microscopic geometric features, and improving the accuracy of reconstruction.
Smart Images
Figure CN121746608B_ABST