A three-dimensional point cloud registration method based on a weighted principal component analysis method and M estimation

The invention discloses a three-dimensional point cloud registration method based on a weighted principal component analysis method and M estimation, and the method comprises the steps of firstly, obtaining a rough and initial conversion relation through a weighted PCA algorithm, and achieving the rough registration of an original point cloud and a target point cloud; then, in order to quickly obtain an accurate rotation translation matrix, using the BP neural network and the two-dimensional moving window for simplifying the number of point clouds of the rotation translation matrix; and finally, adopting a Cauchy function with resistance to noise as an objective function, calculating an accurate alignment relation according to an ICP algorithm, and realizing the accurate registration. According to the method, the time, the space complexity and the algorithm complexity of the registration algorithm can be effectively reduced, and an accurate conversion relation can be obtained for the original point cloud cont