Correspondence Method for Registration of Range Images Using Evolutionary Algorithms

摘要

Registration is a fundamental stage in the 3D reconstruction process, which is used to match two or more images that can be taken in different moments, from different sensors, or from different viewpoints. Genetic algorithms have been applied to the registration of images, because of their ability to solve problems of optimization. In this paper, a procedure is presented to solve the registration problem of multiple view range images using evolutionary algorithms. This procedure is focused, on the problem of obtaining the best correspondence between points through a robust method of search on images that are partially overlapped. This correspondence set will allow us to calculate a rigid transformation that registers the images accurately.