Aiming at the problem existing in 3D object recognition methods based on global features and local features recognition effect are not ideal in the scenes with occlusion and clutter

a 3D point cloud matching algorithm based on point pair feature is proposed. The algorithm uses all the point pair features on the model to complete construction of the global model description

and uses the fast voting scheme to match the model point cloud and scene point cloud locally in the reduced two-dimensional space

so as to restore the global pose of the model in the scene. The experimental results show that the recognition effect of this algorithm is ideal in the scenes with occlusion and clutter.