TempleRAIL / SOGMP

[CoRL 2023] SOGMP++/SOGMP: Stochastic Occupancy Grid Map Prediction in Dynamic Scenes
https://openreview.net/forum?id=fSmkKmWM5Ry
MIT License
20 stars 2 forks source link

About the usage of the dataset. #5

Open daleigehhh opened 5 days ago

daleigehhh commented 5 days ago

Hi, @zzuxzt! Great work, thanks for sharinig it with the community! I am now working on Social-Nav and I am interested in occupancy grid prediction and planning. I noticed your dataset comes from the SCAND dataset but I am little confusing about how to use the npy files provided by you especially the "scan" part. Did you capture the ranges in the ROS bag files directly into your npy files? And I guess the max range is 30 meter? Thank you

zzuxzt commented 3 days ago

Thanks for your interest in our work. The lidar scan .npy file is a numpy array of scan range values. You can directly use python np.load() to read the npy file and obtain the scanning range values, and you can find the scanning bearing values (the angle range and resolution) based on the specific lidar sensor installed on the robot. The maximum range is also based on the specific lidar sensor, but I manually cut it off to 30 meters when I used it.