lidar-camera-placement

KO | EN

gitlite — search

#react #ai #typescript #llm #go #machine-learning #linux #docker #macos #java #windows #awesome #vue #openai

lidar-camera-placement

(★ 21)

[ICRA'24] Influence of Camera-LiDAR Configuration on 3D Object Detection for Autonomous Driving

lidar-camera-placement Latest Version Download

Download Latest Version (.zip)

// repository documentation

Was this content helpful?

★ 0 (0 ratings)

# Camera LiDAR Configuration for 3D Detection
This is the official code for ICRA 2024 "Influence of Camera-LiDAR Configuration on 3D Object Detection for Autonomous Driving". (https://arxiv.org/abs/2310.05245)

## Preparation
Download [CARLA v0.9.10](https://carla-releases.s3.eu-west-3.amazonaws.com/Linux/CARLA_0.9.10.tar.gz) and unzip it under `./carla`. Follow the install instruction of `scenario_runner` of commit [ad71a2c](https://github.com/carla-simulator/scenario_runner/tree/ad71a2c7ed012d735be2b1158fca51b0761ff26b).

We test our repo on Ubuntu 18.04 and 20.04.

## Data Collection

Turn on CARLA as default and then run the following shell script.

`cd ./carla-nuscenes/scripts`

`bash routes_baselines.sh`

LiDAR configurations can be accessed at `./carla-nuscenes/hyperparams`. Camera configurations can be accessed in `./carla-nuscenes/scenario_runner/lidar.py`. The density of Vehicles and Pedestrians can be accessed on `Line 397-422` in `./carla-nuscenes/scenario_runner/srunner/scenarios/route_scenario.py`. The datasets can be accessed at `./carla-nuscenes/dataset`.

Run the following code to create NuScenes-formatted labels.

`cp -r ./carla-nuscenes/dataset/[YOUR LIDAR PLACEMENT]/training/label_2 ./NuScenes_generate/label_2`

`python ./NuScenes_generate/create_test_wide.py`

`python ./NuScenes_generate/create_trainval_wide.py`

Copy the folder `./NuScenes_generate/maps`, `./NuScenes_generate/v1.0-trainval` and `./NuScenes_generate/v1.0-test` to your dataset root. Replace the file `splits.py` in the NuScenes-devkit on your environment with our `./NuScenes_generate/splits.py`, otherwise the NuScenes-devkit can not recognize our dataset.

## Unified Surrogate Metric

The 3d bounding boxes of all 3d objects can be accessed at `./S_MS/routes`. Run the following code to calculate Unified Surrogate Metric (S-MS).

`python ./S_MS/create_camera_angle.py`

`python ./S_MS/evaluate_camera_position.py`

`python ./S_MS/evaluate_lidar_position.py`

## Citation
If you find this work helpful for your research, please kindly consider citing our papers:

```bibtex
@inproceedings{li2024influence,
  title={Influence of Camera-LiDAR configuration on 3D object detection for autonomous driving},
  author={Li, Ye and Hu, Hanjiang and Liu, Zuxin and Xu, Xiaohao and Huang, Xiaonan and Zhao, Ding},
  booktitle={2024 IEEE International Conference on Robotics and Automation (ICRA)},
  pages={9018--9025},
  year={2024},
  organization={IEEE}
}
```

### Reference
> - [Multi-LiDAR-Placement-for-3D-Detection](https://github.com/HanjiangHu/Multi-LiDAR-Placement-for-3D-Detection)
> - [CARLA v0.9.10](https://carla.readthedocs.io/en/0.9.10/)
> - [ScenarioRunner](https://github.com/carla-simulator/scenario_runner)
> - [NuScenes-devkit](https://github.com/nutonomy/nuscenes-devkit)

lidar-camera-placement

lidar-camera-placement Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

lidar-camera-placement

lidar-camera-placement Latest Version Download

Today's picks

Was this content helpful?

Recent Feedback

Download README