Road boundary extraction and vectorization method fusing vehicle-mounted image and point cloud

The invention provides a road boundary extraction and vectorization method fusing a vehicle-mounted image and a point cloud. According to the invention, based on a point cloud super voxel segmentation result, a contraction distance clustering algorithm is adopted to extract and denoise road boundary point clouds; based on the vehicle-mounted image, road boundary pixels are extracted by adopting a semantic segmentation model and a region growing algorithm, and the boundary pixels are further transferred to a world coordinate system; constructing a Snake model, fusing the point cloud and the extraction result of the vehicle-mounted image, and carrying out model solving by utilizing an improved internal force matrix to obtain a road boundary point set; performing vectorization and polyline simplification on the point set to obtain a three-dimensional vector linear road boundary. The road boundary extraction method provided by the invention has relatively strong robustness for three-dimensional road boundaries wi