Road surface pit detection, cleaning and avoidance method based on camera and laser radar

The invention discloses a road surface pit detection, cleaning and avoidance method based on a camera and a laser radar, and the method comprises the following steps: firstly completing the camera-based road surface pit detection based on YOLOv5, carrying out the preprocessing and European clustering recognition of a laser radar point cloud, and obtaining a 3D detection frame; and enabling the 3D detection frame to correspond to the image target detection result through coordinate conversion. Then the 3D detection frame and the 2D detection frame are matched, targets larger than a certain threshold value are fused, and data such as coordinates, the size and the depth of the pothole area are obtained and sent to a downstream cleaning path planning module; for different pothole areas, a sweeper decision is completed based on a finite-state machine model, and whether sweeping time is increased by entering the pothole areas according to an original path or not is judged; or a new path is generated for the deeper