International Journal of Computer Science Issues
The paper proposed a new 3-D measurement point cloud splicing algorithm. The algorithm utilizes registration ideal in model identification technology to realize unbound and accurate splicing of 3-D data. First, sample the overlapping areas in the two 3-D point clouds which need to be spliced. Carry out preprocessing over the sampled point cloud with principal analysis method based on the statistic theory. Through extracting the feature vector that could best indicate the point cloud information, it realizes the dimension reduction for data.