Abstract

The purpose of this research is to examine the suitability of digital photogrammetry for mapping and processing the 3D rock structure of hard rock outcrops using consumer-grade digital cameras and open source software.

For this purpose, the camera settings for photogrammetric image acquisition were first optimized, including camera positions and light paths to achieve sufficient overlapping and to minimize shadowing. The open source software VisualSFM was then used to conduct feature detection, match images, and determine the camera positions. The dense point cloud was constructed using open source algorithms based on the approaches of Patch-based Multi-View Stereo (PMVS) and Clustering Views for Multi-View Stereo (CMVS). The open source software SfM_georef was then used to merge point clouds of different outcrops into a common coordinate system. The accuracy of different smartphone apps for determining the coordinates of Ground Control Points (GCP) using Global Position System (GPS) was tested as an alternative to high accuracy Real Time Kinematic (RTK) GPS.

Rock structures were extracted from the point clouds using the open source software CloudCompare. The statistical computing of the set-based geometrical parameters of discontinuities was conducted in open source software environment R. The elaborated processing methods were tested for five outcrops having different rock types and structural fabrics. Discontinuity parameters estimated using the digital models are compared with traditional contact-based measurements. The results show that photogrammetric modelling and open source computing provides enormous cost, time and safety incentives in standard engineering practice.

1 Introduction

Quantifying the three-dimensional (3D) variability of geological discontinuities representing the spatial organization of a rock fracture network is essential for determining rock block geometries and performing engineering analysis and design in blocky rock masses. These requirements mean that the spatial positions and dimensions of all visible discontinuities should be recorded accurately. Traditional outcrop mapping based on scanline or window mapping using compass and measuring tape can provide the data for an unbiased estimation of set-based statistical parameters such as mean orientation, mean spacing, persistence pattern and their statistical distributions. These parameters represent the fundamental inputs for algorithm-based generations of 3D block systems and discrete fracture network (DFN) models. DFN models are very useful for simulations of fluid flows in fracture network (e.g. Dershowitz et al. 1992) and have also been applied to the design of large-scale rock slopes in open pit mining (e.g. Hoek 2014). The location-dependent 3D rock structure is also important for assessing rock slope failure modes (Goodman and Kieffer 2000), evaluation of rockfall hazards (Dong et al. 2019), or assessment of overbreak and ground behaviour during tunneling (Fekete et al. 2010). Because of the accessibility, safety, or time constraints, practical methods for remotely and efficiently recording of the location-based 3D rock structure are becoming increasingly imperative. In recent years, technological advances in non-contact 3D measurements allows users to choose from numerous solutions (Luhmann et al. 2006). Close-range photogrammetry is one of the fastest growing application fields that generally considers objects in the range of 1 – 200 m and measurement accuracies between 0.1mm and 1cm. The core technology of close-range photogrammetry is known as Structure from Motion (SfM), which uses the position of the camera as it moves through space to estimate X, Y, and Z coordinates for each pixel of the original image (Fig. 1a). Commercial software (e.g. Agisoft PhotoScan, Pix4D, or Bentley ContextCapture) integrates SfM with Multi View Stereo (MVS) algorithms, resulting in greater efficiency and usability. Producing a 3D digital outcrop can require costly hardware and software. However, recent open source software developments, enumerated in this paper, now offer cost-effective means for rock structure characterization.

This content is only available via PDF.
You can access this article if you purchase or spend a download.