一种基于轻量化语义占用网格的无人物流车可行驶区域感知方法及系统
By employing a lightweight semantic occupancy mesh method, combined with onboard cameras and low-line-count LiDAR, unmanned logistics vehicles can now identify unknown obstacles and perceive drivable areas, reducing hardware costs and computational load while improving safety and decision-making quality.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- HONEYCOMB (WUHAN) MICROSYSTEM TECH CO LTD
- Filing Date
- 2026-04-21
- Publication Date
- 2026-07-17
AI Technical Summary
Traditional obstacle detection methods cannot identify unknown categories of objects. Existing occupancy grid methods are computationally intensive and rely on high-line-count LiDAR, making them difficult to deploy on cost-sensitive unmanned logistics vehicles. They also lack optimization for low-speed and narrow areas.
A lightweight semantic occupancy grid method is adopted. Data is collected by vehicle-mounted cameras and low-line-count LiDAR, and spatiotemporal alignment and feature extraction are performed. Image and point cloud features are fused to predict grid occupancy probability and identify semantic labels. Driving areas are generated by combining temporal smoothing processing.
It enables the identification of various unknown obstacles under low-cost hardware conditions, reduces computational load, improves driving safety, provides rich environmental understanding, and optimizes decision-making quality.
Smart Images
Figure CN122116311B_ABST