Closed SvenMuc closed 3 years ago
Hi, @SvenMuc!
Thank you for your interest in RELLIS-3D.
We used Cartographer to generate pose.txt. You can find the tool and configuration here: https://github.com/unmannedlab/cartographer. You can try this script to run on a RELLIS-3D rosbag first.
You can use any other SLAM or odometry estimation method to output the pose of your vehicle corresponding to each lidar scan. LOAM might also be a good choice: https://github.com/HKUST-Aerial-Robotics/A-LOAM
You can find more method here: http://www.cvlibs.net/datasets/kitti/eval_odometry.php
Hope this helps!
Best wishes!
Hi, really interesting approach and tools.
I'm currently trying to extend the dataset by own recordings. But I'm struggling with the pose.txt generation required by the KITTI labeling tool.
Do you have any recommendation how to generate the poses? In your paper you mentioned, that you used a SLAM approach.
Thanks Sven