A review of 3-D reconstruction based on machine vision

Abstract

Reconstructing the three-dimensional (3-D) shape of an object is always a significant study in machine vision. Since it was first proposed by Marr in 1970s，3-D reconstruction has developed a lot but is still limited in many applications. In this paper, various kinds of theories, algorithms and methods to reconstruct 3-D shape are introduced, classified and analyzed. Then, image acquisition methods, computing methods and their achievements and limits are summarized. In the end, some conclusions and advices for machine vision based 3-D reconstruction are presented.