Composite Ski-Resort Registration Method Based on Laser Point Cloud Information

The environment of ski resorts is usually complex and changeable, there are few characteristic objects in the background, which creates many difficulties for registration ski-resort point cloud datasets. However, traditional iterative closest (ICP) algorithm, two points need to have good initial positions, otherwise it easy get caught up local optimizations registration. Aiming at this problem, according topographic features resorts, paper put forward a coarse method based on extraction, matching between feature proposed adjust position clouds. Firstly, common part datasets extracted SIFT algorithm; secondly, Euclidean distance normal vectors used as pairing condition complete datasets; then, pair purified by using included angle vector; finally, process registration, rotation matrix translation vector clouds solved unit quaternion method. Experiments demonstrate that helpful smooth completion subsequent fine process, avoids phenomenon falling into optimization, effectively completes