A novel approach for three-dimensional rock joint surface visualization based on photogrammetric point clouds

Abstract

Joints play a significant role in the behavior of rock masses subjected to stress or deformation. Accurate representations of joint surfaces are crucial for predicting the mechanical behavior of rocks and ensuring the safety of engineering structures. While CT scanning can provide precise joint information, its high cost, limited sample size range, and strict operating environment diminish its practicality for batch measurements of three-dimensional (3D) joint morphology. To address these challenges, this research proposes a cost-effective method for visualizing internal rock joints on a large scale. The proposed method transforms joint image information into 3D point data using image segmentation, feature extraction, digital image processing, and photogrammetry techniques. Based on principal component analysis combined with biharmonic spline interpolation (hereafter referred to as PCA-BSI), a reverse modeling of the 3D joint surface is achieved, enabling visualization of the joint surface. The effectiveness of the PCA-BSI algorithm has been validated through complex examples, showing correlation coefficients close to 1. The reliability of the proposed joint surface reconstruction method was verified by 3D scanning experiments. Furthermore, the joint surface fitting effects of the proposed method were compared to those of global polynomial interpolation and inverse distance weighting methods. The results indicate that the proposed method yielded high-quality 3D visualization of the joint surface, offering an economical and effective alternative for batch visualization of the internal morphology of rock joints.