A local tangent plane distance-based approach to 3D point cloud segmentation via clustering. (May 2023)

• Record Type: Journal Article
• Authors: Chen, Hui; Xie, Tingting; Liang, Man; Liu, Wanquan; Liu, Peter Xiaoping
• Is Part Of: Pattern recognition. Volume 137(2023)
• Publication Date: 2023-05
• Subjects: 3D point cloud -- Plane segmentation -- Tangent distance -- Adaptive clustering; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2023.109307
• Languages: English
• ISSNs: 0031-3203

Abstract

Highlights: A new similarity measure is proposed based on point-to-plane distance. This metric is able to give high and low similarities between coplanar and non-coplanar points, and reconstructing point clouds in 3D space can simplify the segmentation problem. Using the nearest neighbor weighted method to estimate the local density, the obtained density distribution of the reconstructed point cloud is more accurate. In order to overcome the difficulty associated with the selection of DBSCAN parameters due to high sensitivity in the existing work, an adaptive method is developed to cluster reconstructed point clouds. Experiments show that it can achieve the same or even better performance. Abstract: This paper proposes an effective measure for the planar segmentation problem based on the clustering method. It uses the distance from a point to the local plane as a metric to characterize the relationship between data. As a result, the data points of the coplanar have a high similarity to distinguish each plane. A dissimilarity matrix of the input point cloud can be evaluated, and multidimensional scaling analysis is performed to reconstruct the correlation information between data points in the 3D Euclidean space. The obtained reconstructed point cloud shows the separation between different planes. An adaptive DBSCAN clustering method based on density stratification is developed to perform cluster segmentation on the reconstructed point cloud. Experimental results show that the proposed method can effectively solve the over-segmentation problem, and at the same time provide high segmentation accuracy.