Description
Hi,
Thank you for sharing your code. We are aiming to develop path-planning algorithms for robotic navigation in 3D environments (e.g. forest). I read your papers (since 2016) on 3D navigation and found your work interesting. To get the proof-of-concept, I'm trying to replicate path planning in a simple environment (i.e. playpen) using ROS noetic and Gazebo. I used 10 pcd files (recorded using HDL-32E lidar) to create the mesh environment by using the following steps:
1- Filter pcd files using filter_cloud.py (with default params except filter_radius = 0.8) and save as ply files
2- Register ply files using register_scans.py (with default params except filter_radius = 0.8)
3- Combine ply files using combine_scans.py (with default params except filter_radius = 0.8). Moreover, I didn't down sample the combined cloud because the original pcds are very small. Apparently, the combined cloud looks good (see combined_plys.png).
4- Reconstruct the mesh environment using reconstruct.py (with default params except filter_radius = 0.8). The resultant mesh file shows a black rectangle (see front.png). The flipped side show a kind of playpen environment but in bad look (see flipped.png). It is observed that the construction process generate a warning "Extract bad average roots: 21". I can see a very nice mesh file for botanical garden and other environments in your experiments.
It would be a great favor if you can guide me how to fix this issue? Thanks in advance.
Combined Plys
Front Mesh
Flipped Mesh
Original pcds
pcds.zip