A path trajectory-based grid map construction method
By extracting the boundaries of drivable areas through image semantic segmentation and inverse perspective transformation, and combining keyframe filtering and flood filling algorithms to generate raster maps, the high cost and manual annotation problems of existing technologies are solved, and efficient and automated raster map construction is achieved.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHONGQING SHINERAY AGRI MACHINERY
- Filing Date
- 2026-04-28
- Publication Date
- 2026-07-17
AI Technical Summary
Existing raster map construction methods rely on high-precision dense point cloud acquisition, which is costly, difficult to adapt to environmental changes, lacks semantic understanding capabilities, and manual annotation is time-consuming, labor-intensive, and prone to errors.
A path-trajectory-based raster map construction method is adopted. The boundary pixels of the drivable area are extracted through image semantic segmentation, and boundary polygons are generated by combining inverse perspective transformation and keyframe filtering. The raster map is generated by a breadth-first search flood fill algorithm.
It improves the automation and robustness of raster map construction, reduces costs, enhances environmental adaptability and construction accuracy, and reduces human intervention and computational complexity.
Smart Images
Figure CN122408779A_ABST