To solve the problem that the low accuracy and the excessive drift in large distance of the laser odometry of the three-dimensional laser SLAM (simulta-neous localization and mapping) algorithm, the point cloud matchi...
详细信息
ISBN:
(数字)9781665458641
ISBN:
(纸本)9781665458641
To solve the problem that the low accuracy and the excessive drift in large distance of the laser odometry of the three-dimensional laser SLAM (simulta-neous localization and mapping) algorithm, the point cloud matching adopts the multi-metriclinearleastsquareicp (mulls-icp) algorithm. Firstly, the ground segmentation is realized based on the dual-threshold ground filter, then the fusion of IMU and RTK based on error state Kalman filter (ESKF) provides accurate initial value for point cloud matching, Finally, the laser odometry is optimized based on the multi-metriclinearleastsquares icpalgorithm. The experimental results based on the standard datasets KITTI and MVSEC datasets show that compared with the classic ALOAM and LIO-SAM algorithms, the proposed algorithm has higher positioning accuracy and better mapping effect. Compared with LIO-SAM algorithm, the error is reduced by fifty to sixty percent.
暂无评论