prepare_dataset.md

SemanticKITTI

To prepare for SemanticKITTI dataset, please download the KITTI Odometry Dataset (including color, velodyne laser data, and calibration files) and the annotations for Semantic Scene Completion from SemanticKITTI. Put all .zip files under OccFormer/data/SemanticKITTI and unzip these files. Then you should get the following dataset structure:

OccFormer
├── data/
│   ├── SemanticKITTI/
│   │   ├── data_velodyne/
│   │   │   │   ├── velodyne/
│   │   ├── dataset/
│   │   │   ├── sequences
│   │   │   │   ├── 00
│   │   │   │   │   ├── calib.txt
│   │   │   │   │   ├── image_2/
│   │   │   │   │   ├── image_3/
│   │   │   │   │   ├── voxels/
│   │   │   │   ├── 01
│   │   │   │   ├── 02
│   │   │   │   ├── ...
│   │   │   │   ├── 21

Preprocess the annotations for semantic scene completion:

python projects/mmdet3d_plugin/tools/kitti_process/semantic_kitti_preprocess.py --kitti_root data/SemanticKITTI --kitti_preprocess_root data/SemanticKITTI --data_info_path projects/mmdet3d_plugin/tools/kitti_process/semantic-kitti.yaml

NuScenes

Please download nuScenes full dataset v1.0, CAN bus expansion, and nuScenes-lidarseg from the official website. The dataset folder should be organized as follows:

OccFormer
├── data/
│   ├── can_bus/
│   ├── nuscenes/
│   │   ├── maps/
│   │   ├── samples/
│   │   ├── sweeps/
│   │   ├── v1.0-test/
|   |   ├── v1.0-trainval/
|   |   ├── lidarseg
|   |   │   ├──v1.0-trainval/
|   |   │   ├──v1.0-mini/
|   |   │   ├──v1.0-test/
|   |   ├── nuscenes_infos_temporal_train.pkl
|   |   ├── nuscenes_infos_temporal_val.pkl
|   |   ├── nuscenes_infos_temporal_test.pkl

To generate the above data infos, directly download infos or prepare yourself by running:

python tools/create_data.py nuscenes --root-path ./data/nuscenes --out-dir ./data --extra-tag nuscenes --version v1.0 --canbus ./data