Research on a point cloud registration method based on the combination of hash functions and the grey wolf optimizer

Abstract Point cloud registration is the process of determining the optimal rigid body transformation matrix based on the corresponding point pairs between the source point cloud and the target point cloud to align the point clouds. This technology has been widely applied in synchronous localization and mapping, 3D reconstruction, reverse engineering and other fields. However, the existing algorithms often have problems such as low registration accuracy and high computational complexity. This paper proposes a point cloud registration method that combines the hash function with the Grey Wolf Optimizer(GWO). In the rough registration stage, a hash function is first used to preprocess the data in order to efficiently find matching point pairs and calculate the initial registration matrix. Subsequently, the Grey Wolf Optimizer is utilized to optimize the coarse registration matrix of the point cloud. Experiments conducted on several benchmark models (including Bunny, Buddha and Armadillo) show that, compared with traditional algorithms, the method we proposed significantly improves average accuracy and processing speed. This makes it an effective method for point cloud registration.