- LiDAR Projection: This file stores LiDAR points projected onto the image using camera calibration data.
- Object Detection: Only points within YOLO's bounding boxes for the desired objects are kept.
- Outlier Filtering: Filtering removes outliers, which are points that don't truly belong to the objects.
- Bounding Box Adjustment: Two methods are explored for outlier removal:
- Shrinking bounding boxes to ensure only points truly belonging to the objects are considered.
- Applying the Sigma Rule to keep points within 1 or 2 standard deviations of the average distance, based on point distances.
- Focus on Specific Objects: This filtering ensures the data focuses on the specific objects of interest.
git clone https://github.com/Vishalkagade/Camera-Lidar-Sensor-Fusion.git
cd Camera-Lidar-Sensor-Fusion/
pip install -r requirements.txt
Download the 3D KITTI detection dataset from here.
The downloaded data includes:
- Velodyne point clouds (29 GB)
- Training labels of object data set (5 MB)
- Camera calibration matrices of object data set (16 MB)
- Left color images of object data set (12 GB) (For visualization purpose only)
Note- In repository, the required dataset for testing is already added in ./data folder. So for interference, no need to download.But in caseif you want to run it on video, then you need to download the dataset.
python test.py --model yolov8s.pt --img_path data/img --pcd_path data/velodyne --label_path data/label --calib_path data/calib
If you think this work is useful, please give me a star!
If you find any errors or have any suggestions, please contact me (Email: kagadevishal@gmail.com
).
Thank you!