Multi-vision information fusion of a laser scanning system

3D environment reconstruction based on the data of laser radar, which is an important subject in the research and applications of optical 3D sensing, is very important for promotion of optical 3D sensing. This paper has presented an 3D laser scanning system based on 2D laser radar and embedded control unit and demonstrated the 3D model reconstruction from collected range image. The Iterative Closet Point (ICP) algorithm has been used for the registration of 3D geometric data from different viewpoints. The ICP algorithm adopts quaternion expression for computing the rotation matrix R and translation vector t to realize registration. The paper has summed up the choice of the closet point and presented a new solution for the speeding up of the ICP algorithm by introducing five constraint conditions of registration point pair. After the false pair of points being canceled, the number of point pair decreases. The run time is shorten as well.