search

Qutom / BoundaryDetection-Algorithm-for-Point-Cloud-Data

Implement "Novel algorithms for 3D surface point cloud boundary detection and edge reconstruction" using Python
15 stars 1 forks source link

Implementation just for 2D point cloud? #1

Open HowcanoeWang opened 1 year ago

HowcanoeWang commented 1 year ago

Thanks a lot for implementing this algorithm in python for this paper!

When using this code, I got a problem.

According to the demo code in main.py and the figures in this project, did you convert the 3D point cloud to 2D point cloud first (by plane_extractor.extract_plane(pc)), and then run this algorithm on the produced 2D point cloud?

I found the output boundary was only 2D coordiante (x, y, 0), and fit well with the 2D boundary detection

image

When using the latter code in main.py to convert back to 3D (I guess it provides such function for that), the result does not work well on the demo data by paper at all:

image

Qutom commented 1 year ago

Yes, I edited code to find the wall connected to the ceiling of the building. I removed z-axis info in plane_extractor.extract_plane(pc). so It will work just for 2d point cloud

(++ circle class also supports only two dimensions. So expand to 3D you need modify Circle.py, BPD.py and don't use plane_extractor.extract_plane(pc))