Rock joint detection from 3D point clouds based on colour space

To facilitate the automation accuracy of rock joint mapping, a new method based on colour space was proposed for the semi-automatic identification and orientation calculation of rock joints. The developed method in this study comprises four-step: (1) the point colour space and point curvature were calculated based on the point normal and the X-, Y- and Z-coordinates, respectively; (2) the rock joint sets were identified from point clouds based on the difference in point colour space and point curvature; (3) each single rock joint was extracted from the aforementioned joint sets using a density-based spatial clustering of applications with noise (DBSCAN); and (4) the orientation was determined according to the point normals of the fitting planes of the points on each detected rock joint. A dodecahedron was used to demonstrate the procedures of the rock joint detection and orientation calculation, and two outcrop cases were selected to further verify the effectiveness of the proposed method. The results of all cases indicate that the orientation difference between manual measurement and the proposed method was less than 2 degrees. The codes that support the findings of this study have been shared publicly on GitHub (https://github.com/DisDet/ DisDetCIELAB).

Automated summary

A new method based on colour space was proposed for the semi-automatic identification and orientation calculation of rock joints to improve the automation accuracy of rock joint mapping. The method includes four steps: calculating point colour space and point curvature, identifying rock joint sets from point clouds, extracting single rock joints using DBSCAN, and determining the orientation based on the point normals. The effectiveness of the proposed method was verified through examples.