Code for: Spectral analysis and random reconstruction of realistic rock joint surfaces

Published: 13-04-2021| Version 2 | DOI: 10.17632/dygjf6wgbb.2
Shuaihao Zhang ,
Xiang Wang ,
Yingbin Zhang ,
Shi Zuo,
Dejian Li


Research on the mesoscopic morphological characteristics of joint surfaces has been a topic of considerable interest. Based on discrete cosine transform theory, this code presents an improved spectral analysis workflow for quantitatively characterizing and reconstructing the 3D mesoscopic morphology of rock joint surfaces. First, 3D cloud point data of realistic rock joint surfaces are obtained via photogrammetry scanning. Next, the joint surfaces are segmented and transformed into standard grid data sets, which are regarded as discrete time-domain signals. Based on the discrete cosine transform algorithm, it is possible to convert such a discrete signal in the time domain into a discrete spectrum in the frequency domain. Then, the normalized amplitudes of the modified spectrum are defined as joint morphology descriptors, which are found to be applicable for the characterization and reconstruction of joint surfaces. Accordingly, the roughness coefficients of 310 measured 3D joint samples and 21,000 virtual joint surfaces are evaluated. It is found that joint surfaces with specific roughnesses can be reconstructed. This code provides a useful tool for the rapid random reconstruction of joint surfaces, which is beneficial for providing virtual test samples for numerical modelling. For more detailed instructions, see the readme file in the compressed package.


Steps to reproduce

For more detailed instructions, see the readme file in the compressed package.