Skip to content

About

[CVPR 2024] Collaborative Semantic Occupancy Prediction with Hybrid Feature Fusion in Connected Automated Vehicles- data generator

Resources

Stars

2 stars

Watchers

1 watching

Forks

Repository files navigation

CoHFF Data Generator for OpenCOOD

This repository provides an extended data generator for:

Collaborative Semantic Occupancy Prediction with Hybrid Feature Fusion (CoHFF)
in Connected Automated Vehicles (CAVs), based on OpenCOOD.

📄 Project Page: https://rruisong.github.io/publications/CoHFF/


🛠️ Setup Instructions

✅ Step 1: Clone OpenCOOD Base

git clone https://github.com/DerrickXuNu/OpenCOOD

✅ Step 2: Clone This Data Generator

git clone https://github.com/chrisyan/opv2v_datagenerator_CoHFF.git

🚗 Configuration Options

You can generate data for different multi-agent LiDAR setups. Choose one of the following:


Option 1: 4 Semantic LiDARs

  1. Copy configuration and sensor files:
cp opv2v_datagenerator_CoHFF/logreplay/hypes_yaml/replay_4.yaml OpenCOOD/logreplay/hypes_yaml/
cp opv2v_datagenerator_CoHFF/logreplay/sensors/semantic_4/semantic_lidar.py OpenCOOD/logreplay/sensors/
cp opv2v_datagenerator_CoHFF/logreplay/sensors/semantic_4/sensor_manager.py OpenCOOD/logreplay/sensors/
  1. Modify config path in scenarios_manager.py:

Open the file:

OpenCOOD/logreplay/scenario/scenarios_manager.py

Change line 73 from:

scene_params = load_yaml('../hypes_yaml/replay.yaml')

to:

scene_params = load_yaml('../hypes_yaml/replay_4.yaml')

Option 2: 18 Semantic LiDARs

  1. Copy configuration and sensor files:
cp opv2v_datagenerator_CoHFF/logreplay/hypes_yaml/replay_18.yaml OpenCOOD/logreplay/hypes_yaml/
cp opv2v_datagenerator_CoHFF/logreplay/sensors/semantic_18/semantic_lidar.py OpenCOOD/logreplay/sensors/
cp opv2v_datagenerator_CoHFF/logreplay/sensors/semantic_18/sensor_manager.py OpenCOOD/logreplay/sensors/
  1. Modify config path in scenarios_manager.py:

Open the file:

OpenCOOD/logreplay/scenario/scenarios_manager.py

Change line 73 from:

scene_params = load_yaml('../hypes_yaml/replay.yaml')

to:

scene_params = load_yaml('../hypes_yaml/replay_18.yaml')

▶️ Step 5: Run the Data Collection

cd OpenCOOD/logreplay/scenario
python3 scenarios_manager.py

This will collect the replayed semantic LiDAR data in a format usable for CoHFF training and evaluation.


📌 Notes

  • Ensure OpenCOOD is correctly installed. Follow their installation guide.
  • You can customize the number of agents or sensor configurations by creating your own YAML and sensor manager classes.

📚 Citation

If you find this data generator useful, please cite:

@INPROCEEDINGS{10658094,
  author={Song, Rui and Liang, Chenwei and Cao, Hu and Yan, Zhiran and Zimmer, Walter and Gross, Markus and Festag, Andreas and Knoll, Alois},
  booktitle={2024 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)}, 
  title={Collaborative Semantic Occupancy Prediction with Hybrid Feature Fusion in Connected Automated Vehicles}, 
  year={2024},
  volume={},
  number={},
  pages={17996-18006},
  keywords={Visualization;Solid modeling;Three-dimensional displays;Roads;Semantics;Collaboration;Object detection;Collaborative Perception;Cooperative Perception;Semantic Occupancy Prediction;Semantic Scene Completion;Autonomous Driving;Multi-agent Systems;Robotics},
  doi={10.1109/CVPR52733.2024.01704}}

About

[CVPR 2024] Collaborative Semantic Occupancy Prediction with Hybrid Feature Fusion in Connected Automated Vehicles- data generator

Resources

Stars

2 stars

Watchers

1 watching

Forks

Releases

Packages

Used by

Contributors

Languages