Registering two 3D point clouds

Question

I have one point cloud that describe an object (let's say, a blob of clay). I have another point cloud that describes the same object, but a portion of the object has changed (i.e, let's say that someone ripped off a small piece of clay from the blob).

I'm trying to figure out how to register the two point clouds, so that I can figure which portions of the 1st object are still intact, and which are missing.

I'm very new to registration techniques, so any ideas on how to approach this problem would be appreciated.

Answer 1

This 1990's algorithm, called iterative closest point, is a good start:

Besl, Paul J., and Neil D. McKay. "Method for registration of 3-D shapes." Robotics-DL. International Society for Optics and Photonics, 1992.

It has been cited quite frequently since then (cf. Google Scholar).

Answer 2

Since you know that point clouds are identical, calculate few invariants and align them. For exaple, calculate centroids so you know the translation. Then you can calculate few eigen vectors so you know rotations etc.