摘要

In this paper, we describe a simple, low-cost and compact three-dimensional reconstruction system based on monocular vision and propose an algorithm that can be used to obtain an object's three-dimensional shape. This method requires very little hardware besides a web camera, a hand-held linear red laser, and three spatial background boards that are pairwise orthogonal, such as the corner of a room. The object to be reconstructed is placed on the background boards with the camera facing the object. The operator holds the linear laser, and slowly and uniformly scans the object from top to bottom or vice versa. The three-dimensional object information can be extracted from a collection of images. The most innovative part of this algorithm is that the camera does not require calibration, and the background boards do not need to be marked or annotated. Experimental results show this 3D reconstruction method is robust, effective, and precise.