Point cloud registration method based on improved particle swarm optimization

The invention provides a point cloud registration method based on an improved particle swarm algorithm, and belongs to the technical field of electronic information. A part point cloud obtained through three-dimensional equipment scanning is registered with a part theoretical mathematical model, an improved particle swarm iterative optimization algorithm based on decoupling is used, three rotation matrix parameters and three translation matrix parameters are used as six dimensions of particles, a target function is optimally set by adopting a 3 sigma criterion, and a part theoretical mathematical model is obtained. The intelligent matching of the three-dimensional measurement point cloud and the theoretical mathematical model of the part is realized, and the accurate pose of the matched part relative to the three-dimensional scanning equipment is obtained. According to the method, the global features and the local optimum of the parts are considered at the same time, the point cloud registration time is effec