PointCloudRegistration-py

Testing 4 level point cloud registration

1. filtering (remove noise)
2. down sampling (reduce point)
3. global registration (rough registration)
4. local refinement (elaborate registration)

main dependencies

• tkinter (user interface)
• panda3d (scene graph)
• open3d (point cloud)

Usage

Screenshot

View

mesh	point cloud	down sampled point cloud

In this screenshot, processing means down sampling and filtering. We just do 5 mm scale down sampling at now.

initial	global registration result	local registration result

We support point cloud registration by right buttons.
global point cloud registration -> using feature point ransac (open3d global registration)
local point cloud registration -> open3d g-icp

And we support changing to custom source or target by menu (just only obj file...)

TODO

다양한 PointCloud Registration Option을 넣을 수 있는 User Interface 추가