Point cloud simplification is concerned with reducing the number of redundant points and preserving geometric features, so as to provide a better representation of the underlying surface. In early research, many researchers focused on the moving least squares (MLS) method, volume data, and iterative simplification. MLS is used to construct local surfaces implicitly [3,4], and points are projected to the surface for down sampling. Kobbelt et al [5] simplified point clouds by extracting feature-sensitive surfaces based on volume data. Lipman et al [6] proposed a locally optimal projection (LOP) operator and applied it to raw scanned data with complex shapes. Huang et al [7] developed a weighted locally optimal projection (WLOP) operator based on LOP, which has proven to be less sensitive to noise and has the advantage of producing an evenly distributed point cloud. To reduce the computational complexity of WLOP, Yang et al [8] focused on the decomposition of a point cloud and created multiple output results by Measurement Science and Technology