A review of range image registration methods with accuracy evaluation

Three-dimensional scanning technology allows the acquisition of a geometric model for real-world surfaces. Most of the acquisition systems are limited to reconstruct a partial view of the object obtaining in blind areas and occlusions, while in most applications a full reconstruction is required. Many authors have proposed techniques to fuse 3D surfaces by determining the motion between the different views. Although the motion between these views is usually unknown, it can be computed by means of registration algorithms. In recent years, some methods have been presented: (a) iterative closest point (ICP); (b) Method of Chen; (c) signed distance fields; and (d) genetic algorithms, among others. In this paper, we reviewed the fine Range Image Registration algorithms, and a new evolutionary algorithm which marries simulated annealing with genetic algorithms (GAEA) is introduced to solve this problem. The algorithm avoids the premature convergence problem existed in genetic algorithms, enhances the globe convergence, and improves the convergence velocity. Algorithm is implemented in MATLAB because MATLAB guarantees an easy implementation. Real scanned objects are used to take into account the accuracy of the GAEA. Our experimental results showed that the introduced GAEA provided better solution.