README.md

Lidar_Quadtree_Segmentation

1. get LiDAR data (Ouster OS1 64 or Velodyne VLP 16, or any 3d 360 degree LiDAR)
2. subscribe '/points_raw' topic
3. saving Z axis information
4. 3d Data -> 2d Data ( project onto Road Plane that was estimated by RANSAC )
5. point cloud data segmentation, Using quadtree
6. clustering
7. 2d -> 3d as bounding boxes, Using Z axis information

Hardware

1. HYUNDAI i30
2. Ouster OS1 64 channel LiDAR

Input topic

• '/points_raw'

Output topic

• '/detected_boxes'
• '/projected_cloud'
• '/quad_cluster'

Run

$ roslaunch lidar_detect quadtree_seg_launch.launch --screen 

result

1. segmentation

2. 3D object detection

3. 3D view with PointCloud2(3d pointcloud data)