KITTI-360

We use Semantic DSP Map in work 1) "Particle-based Instance-aware
Semantic Occupancy Mapping in Dynamic Environments" for local
mapping and accumulate the voxels to get a global map. Localization
data comes from ORB-SLAM2 data in 2) kitti360Scripts. Depth is
generated by 3) SGM. 2D segmentation is generated by 4) CMNext .

1) https://github.com/tud-amr/semantic_dsp_map
2) https://github.com/autonomousvision/kitti360Scripts
3) https://github.com/dhernandez0/sgm
4) https://github.com/jamycheung/DELIVER

Voxel resolution: 0.1 m. Maximum particle number in each voxel: 8.
Local map size 51.2 x 51.2 x 25.6 m.

@misc{chen2024particlebasedinstanceawaresemanticoccupancy,
title={Particle-based Instance-aware Semantic Occupancy Mapping
in Dynamic Environments},
author={Gang Chen and Zhaoying Wang and Wei Dong and Javier
Alonso-Mora},
year={2024},
eprint={2409.11975},
archivePrefix={arXiv},
primaryClass={cs.RO},
url={https://arxiv.org/abs/2409.11975},
}