PRBonn / lidar-bonnetal

Semantic and Instance Segmentation of LiDAR point clouds for autonomous driving
http://semantic-kitti.org
MIT License
961 stars 206 forks source link

point de-skewing #117

Closed smileyr99 closed 3 months ago

smileyr99 commented 4 months ago

Hello, first of all, thank you for providing a great paper and code.

I have a question regarding point de-skewing. According to the paper, the de-skewing process is performed to correct the rolling shutter effect, but I could not find this process in the code. Can you explain in detail how the distortion correction process mentioned in the paper is carried out? I would appreciate it if you could provide some information regarding this.

Thank you.

jbehley commented 3 months ago

The deskewing or accounting for the motion of the sensor while scanning the environment has already been performed by the authors of KITTI using the car-grade inertial measurement unit (IMU). Therefore, we rely on the already deskewed point clouds of the KITTI odometry benchmark.

Please see the implementation in KISS-ICP (https://github.com/PRBonn/kiss-icp/blob/main/python/kiss_icp/kiss_icp.py#L45) for a way to deskew the scan based on an estimate of the motion.