Advanced science.  Applied technology.

Search

Integrating SwRI Expertise in Physical Analog Modeling and Digital Photogrammetry for Improved Geologic Analysis, 15-R6360

Principal Investigators
Danielle Wyrick
Inclusive Dates 
05/15/23 to 09/15/23

Background

A key challenge in structural geology and geomechanics is the ability to accurately model the evolution of complex geologic structures in a reproducible and efficient manner. Physical analog models are particularly well suited to simulating the nucleation and evolution three-dimensional (3D) structural complexity and discontinuous deformation in geologic media, and SwRI geoscientists have been simulating geologic deformation with physical analog models for over 25 years. Modern monitoring techniques such as laser scanning and dynamic structured light imaging allow quantitative analysis of analog model deformation in 3D at sub-mm resolution, but are typically not well adapted to changes in physical model configurations (e.g., variable model sizes and analog materials). Further, these techniques can produce data artifacts that mask small-scale deformation. Digital photogrammetry is a flexible and efficient method for supplementing field studies and this method has been successfully adopted at SwRI to supplement field-based geologic analyses. The focus of this project was to combine and integrate SwRI expertise in digital photogrammetry and physical analog modeling of geologic processes.

Approach

Phase 1 focused on (i) reconstructing static objects and comparing physical vs. digital measurements of object dimensions, and (ii) testing the effects of laboratory lighting, camera settings, and modeling material properties on photogrammetric reconstruction accuracy and reproducibility. Six physical analog models aimed at simulating deformation were run during Phase 2. Multiple model configurations were tested, including extensional, contractional, and strike-slip setups, as well as a range of analog materials (sand, clay), and variable lighting conditions. Phase 3 focused on post-processing, data visualization, and quantitative structural analysis. Photogrammetric point-clouds were post-processed to remove geometric artifacts from reconstructions and “cleaned” point clouds were used to derive digital elevation models, slope and aspect maps, and difference maps that highlight deformation induced changes. In addition, overhead images were used to capture patterns of deformation in the horizontal (x-y) plane.

Accomplishments

A total of 81 photogrammetric reconstructions were generated, with photos acquired at 0.3 to 1.1 m from model surfaces. Ground pixel resolutions of 0.72 to 0.048 mm were achieved, with higher resolutions achieved later in the project as laboratory protocols were adjusted and refined. Diffuse lighting conditions provided highest accuracy reconstructions, with brightly lit or flash photography resulting in higher model uncertainties. Low reflectivity granular materials (e.g., volcanic ash) yielded accurate digital reconstruction results with few errors whereas reflective material (e.g., glass bead) models typically suffered from high errors and reconstruction gaps. Material color did not directly affect photogrammetric accuracy. Sprinkling of low reflectivity granular material on reflective wet clay models improved reconstruction results, with higher densities of grain coatings typically yielding lower point-cloud errors. Project experiments allowed optimal conditions and best-practice guidelines for photogrammetric reconstruction of physical analog models to be established. As a result, physical analog modeling and quantitative structural analysis capabilities at SwRI have been greatly enhanced.