Towards semi-automatic rock mass discontinuity orientation and set analysis from 3D point clouds. (June 2017)

• Record Type: Journal Article
• Authors: Guo, Jiateng; Liu, Shanjun; Zhang, Peina; Wu, Lixin; Zhou, Wenhui; Yu, Yinan
• Is Part Of: Computers & geosciences. Volume 103(2017)
• Page Start: 164
• Page End: 172
• Publication Date: 2017-06
• Subjects: 3D point cloud -- Rock mass discontinuity -- Point segmentation -- Cluster analysis -- Firefly algorithm (FA) -- Fuzzy c-means (FCM); Environmental policy -- Periodicals; 550.5
• DOI: 10.1016/j.cageo.2017.03.017
• Languages: English
• ISSNs: 0098-3004

Abstract

Abstract: Obtaining accurate information on rock mass discontinuities for deformation analysis and the evaluation of rock mass stability is important. Obtaining measurements for high and steep zones with the traditional compass method is difficult. Photogrammetry, three-dimensional (3D) laser scanning and other remote sensing methods have gradually become mainstream methods. In this study, a method that is based on a 3D point cloud is proposed to semi-automatically extract rock mass structural plane information. The original data are pre-treated prior to segmentation by removing outlier points. The next step is to segment the point cloud into different point subsets. Various parameters, such as the normal, dip/direction and dip, can be calculated for each point subset after obtaining the equation of the best fit plane for the relevant point subset. A cluster analysis (a point subset that satisfies some conditions and thus forms a cluster) is performed based on the normal vectors by introducing the firefly algorithm (FA) and the fuzzy c-means (FCM) algorithm. Finally, clusters that belong to the same discontinuity sets are merged and coloured for visualization purposes. A prototype system is developed based on this method to extract the points of the rock discontinuity from a 3D point cloud. A comparison with existing software shows that this method is feasible. This method can provide a reference for rock mechanics, 3D geological modelling and other related fields. Highlights: New method for semi-automatic rock mass discontinuity extraction from point clouds. Firefly algorithm (FA) was introduced to discontinuity extraction from point clouds. Octree was used to automatically segment points to multi-scale subsets efficiently. The proposed method can handle more than two million points with high efficiency.