Unstructured Road Slope Recognition Based on Improved RANSAC Algorithm

Abstract

Unstructured road scenes usually have bumpy and undulating features, and there are often obstacles such as potholes and stones on the ground. At the same time, the lidar point cloud with low beams is relatively sparse, which easily affects the accuracy of slope detection results. In response to these problems, this paper proposes a real-time detection method for unstructured road slope based on the improved RANSAC algorithm. Use inertial navigation to obtain vehicle motion information, fuse historical frame point clouds to enrich the point cloud density, then perform multi-scale rasterization on the point cloud, and then use the improved RANSAC algorithm to fit the point cloud, and finally get the slope of the road. Experimental results verify that the algorithm can improve detection accuracy, real-time performance, and effectiveness.