Research on scattered points cloud denoising algorithm

The raw data of cloud produced by 3D scanning tools contains additive noise from various resources. We study the related research at home and abroad. Aiming at the problem that the bilateral filtering algorithm can´t well denoise the points data in the flat region and save the model geometric feature at the same time, this paper introduces the bilateral filter in points g algorithm´s principle and its advantage and disadvantage. Then we improve the search method of any points´ k neighbors of K-D tree and put forward a points cloud denoising algorithm based on feature. This algorithm absorbs the advantages of the bilateral filtering algorithm in image processing and the plane projection algorithm, which can ideally smooth the flat region and can keep the model geometric feature of detail region. At last, we prove the correctness and effectiveness of this paper´s denoising algorithm by the experiments.