README.md

Lidar_Openpcdet_ST3D

Introduction

Recently, I've been working with my own dataset. I realized many might be unfamiliar with how to utilize the advanced LiDAR open source tool, Openpcdet, to train their custom datasets. Hence, I've documented the steps I followed. Regardless of the type of LiDAR you use, ensure you have data in the format (x,y,z,intensity). The subsequent conversions are straightforward, ranging from formats like .txt, .pcd, .bin, to .npy.

Progress

• ✅ Build a dataloader and import your own data
• 📝 Training my own dataset (TODO)
• 📝 Test my own dataset (TODO)

1. Dataloader

1.1 Prepare Your Own LiDAR Data

ST3D/Openpcdet
├── data
│   ├── mydata
│   │   │── lidar
│   │   │── label

│ │ │── lidar

• lidar information

│ │ │── label

• label information

• For lidar annotation tools please refer this : https://github.com/Leozyc-waseda/awesome-lidar-labeling

1.2 Align the coordinates of your own dataset with openpcdet

        # My dataset coordinates are:
        # - x pointing to the right
        # - y pointing to the front
        # - z pointing up
        # Openpcdet Normative coordinates are:
        # - x pointing foreward
        # - y pointings to the left
        # - z pointing to the top
        # So a transformation is required to the match the normative coordinates
        points = points[:, [1, 0, 2]] # switch x and y
        points[:, 1] = - points[:, 1] # revert y axis

1.3 Modify & create the three configuration files

1.3.1 ~/pcdet/datasets/__init__.py

Lidar_Openpcdet_ST3D/ST3D/pcdet/datasets/__init__.py

Line 9 in 365ac6b

from .mydata.mydata_dataset import mydataDataset

Lidar_Openpcdet_ST3D/ST3D/pcdet/datasets/__init__.py

Line 20 in 365ac6b

'mydataDataset':mydataDataset

Feel free to replace mydata with your own dataset name.

1.3.2 ~/pcdet/datasets/mydata/mydata_dataset.py

• Configuration about your own dataset https://github.com/Leozyc-waseda/Lidar_Openpcdet_ST3D/blob/365ac6b57906a01aeac27bfbdd8f1df29b5c2744/ST3D/pcdet/datasets/mydata/mydata_dataset.py

• Get your Lidar path and annotation path

  Lidar_Openpcdet_ST3D/ST3D/pcdet/datasets/mydata/mydata_dataset.py

  Line 88 in 365ac6b

  pcd_path = one_road_path / one_folder #/home/algo-4/work/ST3D/data/mydata/lidar

  Lidar_Openpcdet_ST3D/ST3D/pcdet/datasets/mydata/mydata_dataset.py

  Line 91 in 365ac6b

  label_path = one_road_path / one_folder # #/home/algo-4/work/ST3D/data/mydata/label

• Get all annotation labels informations，Regarding the annotation information, you need to change it to your own annotation information. There is only car in my json file. Since I haven't used it for training, I haven't tested this part yet.

  Lidar_Openpcdet_ST3D/ST3D/pcdet/datasets/mydata/mydata_dataset.py

  Lines 140 to 146 in 365ac6b

  	single_objects_label_info['name'] = np.array([label['obj_type'] for label in labels]) # ['Car']
  	single_objects_label_info['box_center'] = np.array([[label['psr']['scale']['x'], label['psr']['scale']['y'],label['psr']['scale']['z']] for label in labels]) #single_objects_label_info['box_center'] [[4.44983645 1.6537832 1.49163282]]
  	#print("single_objects_label_info['box_center']",single_objects_label_info['box_center'])
  	single_objects_label_info['box_size'] = np.array([[label['psr']['position']['x'],label['psr']['position']['y'],label['psr']['position']['z']] for label in labels])
  	single_objects_label_info['box_rotation'] = np.array([[label['psr']['rotation']['x'],label['psr']['rotation']['y'],label['psr']['rotation']['z']] for label in labels])
  	single_objects_label_info['tracker_id'] = np.array([ label['obj_id'] for label in labels])

1.3.3 ~/tools/cfgs/dataset_configs/mydata_dataset.yaml

• Basically copy from kitti_dataset.yaml

https://github.com/Leozyc-waseda/Lidar_Openpcdet_ST3D/blob/365ac6b57906a01aeac27bfbdd8f1df29b5c2744/ST3D/tools/cfgs/dataset_configs/mydata_dataset.yaml

• change here

  Lidar_Openpcdet_ST3D/ST3D/tools/cfgs/dataset_configs/mydata_dataset.yaml

  Lines 1 to 2 in 365ac6b

  	DATASET: 'mydataDataset'
  	DATA_PATH: '../data/mydata'

1.4 Generate the data infos by running the following command:

python -m pcdet.datasets.mydata.mydata_dataset create_mydata_infos tools/cfgs/dataset_configs/mydata_dataset.yaml

2. Training my own dataset (📝 TODO)
3. Test my own dataset (📝 TODO)

---

📜 License

This project is licensed under the MIT License.

For more details, see LICENSE file.