Point cloud registration method based on distribution

The invention discloses a point cloud registration method based on distribution. The method is mainly composed of a point cloud feature extraction module, an attention overlapping part identification module and a point cloud distributed registration module, and is used in a point cloud registration task with partial overlapping, wherein the point cloud feature extraction module uses an RRI algorithm to extract a point cloud rotation invariant descriptor; the attention identification module is composed of a GNN, a multi-head self-attention module and a multi-head cross-attention module; geometrical features and global features of point clouds are learned, and overlapping points of source point clouds and target point clouds are effectively identified; and the point cloud distributed registration module is composed of three layers of full connection, GMM distribution parameters of the point clouds are obtained and used for point cloud registration calculation, and a rotation matrix and a translation matrix of t