A fully automatic-image-based approach to quantifying the geological strength index of underground rock mass

Sen Yang, Shimin Liu, Nong Zhang, Guichen Li, Jie Zhang

Research output: Contribution to journalArticlepeer-review

Abstract

The conventional discontinuity survey process in the mining industry is known to be a time-consuming one and it is technically challenging due to the limited accessibility of fresh rock exposures. A rapid and robust rock mass property quantification system is desirable for rock structure design during mining operations. In this work we develop an image-based and fully automatic rock mass Geological Strength Index (GSI) rating system. The proposed method involves a series of novel algorithms to quantify the GSI rating based on data recovered from digital images. The proposed GSI system includes both Structure Rating (SR) and Joint Condition Digital Imaging (JCDI) to represent the bulk rock and discontinuity surface conditions of the rock mass. Local histogram equalization and self-adaptive gamma correction were introduced into the pre-processing of the rock face images. Compared to conventional histogram equalization, local histogram equalization can effectively restrain the uneven distribution of brightness often present. Self-adaptive gamma correction based on the image properties automatically enlarges the difference between discontinuity areas and intact rock. A novel algorithm combining region growing and the Hough transform is proposed for the automatic extraction of discontinuity areas. Laboratory and field tests demonstrated that the algorithm possesses an advantage over existing methods with regard to better noise suppression and that it can yield reasonable results for images taken in poor photography conditions. Discontinuities were characterized using a novel algorithm comprising area thinning, skeleton linking, spurs removal, and sampling. Using this algorithm, four parameters of the discontinuity can be quantified: length, orientation, separation width, and JRC value. The proposed approach was validated by two field-based case studies.

Original languageEnglish (US)
Article number104585
JournalInternational Journal of Rock Mechanics and Mining Sciences
Volume140
DOIs
StatePublished - Apr 2021

All Science Journal Classification (ASJC) codes

  • Geotechnical Engineering and Engineering Geology

Fingerprint Dive into the research topics of 'A fully automatic-image-based approach to quantifying the geological strength index of underground rock mass'. Together they form a unique fingerprint.

Cite this