Initial Position and Orientation Estimation of Scan Matching Based on the Height Feature Map

In order to solve the problem that the uniform speed model of lidar odometer can't accurately predict the location at variable speed, the method based on height feature map is proposed in this paper, which extracts co-visible regions between adjacent frames and matches them in feature grid maps to obtain initial position and orientation. By testing it on the kitti data and compared it with F-LOAM, this method shows better accuracy in trajectory and improves the stability of the algorithm at low iteration.