The scattered point cloud denoising method based on surfacelet transform of 3D context

This paper proposed a fast denoising disposal method for scattered point cloud. The best approximation plane is determined by application of least-squares method and a new three-dimensional coordinates is established by the law vector of the plane as Z-axis, Scattered point will be coordinate transformation to new coordinate system; to conversion of scattered data process surfacelet transform and use improved 3D context model to estimate coefficient threshold of transformed different scales. According to threshold, the coefficient is filtered for denoising aim and the scattered point cloud convergence for a stable 3-D digital model. The experiment results show that noise is significant suppressed after denoising fairing data. Compared with other methods, processed data is more fairing and this method better retains detail features of the scattered points of 3D shape. This method makes full use of the relationship of 3D point cloud and provides the necessary preconditions for further three-dimensional reconstruction work.