如何逐步合并3D网格

Question

I'm using xbox kinect as my sensor and I'm trying to create a 3d map from it. Currently I am able to generate point clouds from its depth sensor. Once we get the point cloud, it can be converted to a triangle mesh. The problem is we will get multiple 3d meshes and we need to somehow detect the overlap among all the meshes and create one single mesh.

How do I detect the overlap and combine all the meshes into one mesh? Are there good algorithms that I can use?

Answer 1

My suggestion is to convert this map into gradient map. Wherever there is a high gradient pixel, it means it is a boundary.

Then, separate the map by the boundaries you calculate before.

Now you should have good separation between objects.

Finally, you can apply your mesh algorithms on the separated maps.

(Hope I've understood you well)