3D reconstruction and learning from LIDAR Sensor

Automatic 3D reconstruction and learning from sparse 2.5D LIDAR images. (LIDAR, 3D reconstruction, 3D segmentation, 3D classification, depth view registration, ICP)

3D reconstruction from sparse 2.5D depth map from LIDAR. Reconstructure 3D mesh is segmented and classified as part of interest for measurement and evaluation.

LIDAR configuration and arrangement on the measuring robotic platform. Raw results are depth maps.
Results of trunk and branch modeling on synthetic and real data.
  1. Reconstruction; 2. Vertical Segmentation; 3. Horizontal Segmentation; 4. Trunk and branch modeling using cylinder.