/Object-Detection-based-Lidar-and-camera

Object detection based lidar-and-camera

Primary LanguageJupyter Notebook

Object-Detection-based-Lidar-and-camera

Install mmlab

https://mmdetection3d.readthedocs.io/en/stable/getting_started.html#

Real envs

1. Connect lidar and camera

roslaunch velodyne_camera connect_lidar_camera.launch

2. get velodyne data using python

python src/detection_pkg/MMLAB_based/point_pillar/src/get_velodyne_data.py

3. Run PointPillar

python src/detection_pkg/MMLAB_based/point_pillar/src/real_lidar_pillars.py

or 

python src/detection_pkg/MMLAB_based/real_lidar_pillars.py

or 

python src/detection_pkg/MMLAB_based/rosbag_pillars.py

Gazebp envs

1. Run gazebo world

roslaunch mecanum_robot_gazebo mecanum_velodyne.launch

2. Run Second

python src/detection_pkg/second.pytorch/test.py

3. Run Second

python src/detection_pkg/second.pytorch/real_lidar_pillars.py

utility

  • kitti data to rosbag 
    kitti2bag -t {folder-name} -r {folder-number} raw_synced .
  • play rosbag 
    rosbag play {rosbag-name}.bag --clock --pause --loop