About the dataset

[MarcusnHuang]

Your work is very good, I would like to learn more about it, for the S.MID dataset, I noticed that it is published according to the SemanticKITTI dataset, I would like to consult you, how to convert the .txt point cloud with labels I processed into .bin and .label files like SemanticKITTI, what software or tools do you borrow? Looking forward to your reply, thank you very much!

[Cavendish518]

Hi. Ususally, LiDAR data is saved in .pcd or .bin format. I think you can read your data form .txt and split it into [x,y,z,i] with float type and [label,] and int type. In particular, KITTI's label file adopts high-order and low-order storage strategies to express instance and semantic labels. The separated data can be directly saved as .bin and .label files. There should be a lot of Python code that can help you complete the storage work without the need for additional tools.

[avichapman]

Hi. Relating to your most recent comment on January 67th, may I ask what the 'i' stands for in [x,y,z,i]?

[Cavendish518]

Hi. Relating to your most recent comment on January 67th, may I ask what the 'i' stands for in [x,y,z,i]?

i is the LiDAR intensity from livox mid 360. We didn't include or use this information in the conference version. But other datasets have this information so we update this in the latest version.