A variant of point-to-plane registration including cycle minimization

3D models are very important in many industrial and scientific applications. Most part of commercial sensors obtain only a partial acquisition of the object, so that a set of views are required to build a complete model of the object. Although the motion between these views is usually unknown, it can be computed by means of registration algorithms. A survey of most important techniques is presented in this paper, in which they have been classified into coarse and fine registration and compared in terms of the number of views aligned at every step, the accuracy and the robustness against outliers. The second part of the article presents an improvement of point-to-plane registration, which includes the determination of cycles in a sequence of views with the aim of minimizing the propagation error or drift.