search

RobustFieldAutonomyLab / LeGO-LOAM

LeGO-LOAM: Lightweight and Ground-Optimized Lidar Odometry and Mapping on Variable Terrain
BSD 3-Clause "New" or "Revised" License
2.42k stars 1.13k forks source link

Running Kitti data set produces serious drift #226

Closed Mushan-fu closed 3 years ago

Mushan-fu commented 3 years ago

When I run the Kitti data set in the paper with the source code, the effect of building the map is very poor, and it can't even be similar to the ground truth of the data set. It's completely another map, and the drift is extremely serious. When I use the 09 sequence data set, the drift is also serious. Only when the topic of IMU is removed can the better effect be achieved, otherwise, the paper can not be built. What is the reason for this? Thank you。

stale[bot] commented 3 years ago

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.