Effects of Analog Modeling Materials on Topographic Photogrammetry (SfM) Reconstructions

Abstract

Accurate topographic data are essential for quantitative structural analysis, both in natural settings and in the laboratory. The selection of modeling materials (with appropriate rheological properties) is known to be fundamental for the success of scaled physical analog experiments. However, the optical properties of analog materials and their impact on the reliability and precision of high-resolution topographic reconstructions have not (to our knowledge) previously been assessed. Here we evaluate the effects of material composition, color, and grain size on Structure-from-Motion (SfM) photogrammetry reconstruction efficacy for deformed and undeformed model configurations in the laboratory. Image collections for photogrammetry are acquired from multiple camera positions with a handheld digital camera, and reconstructions are registered using ground control points in a local coordinate system. Static experiments show that low reflectivity granular materials (e.g., silica sand, volcanic ash, pumice, and Al₂O₃) yield relatively reliable photogrammetry data for a wide range of grain sizes (44–2,400 μm) but larger grain sizes (≥250 μm) provide more robust results. Reflective materials (e.g., glass beads, wet clay) yield less reliable point-clouds but the addition of low-reflectivity granular materials (e.g., Al₂O₃ grains) on the surface of wet clay improves reconstruction results, with higher grain densities typically yielding lower point-cloud residuals. SfM-photogrammetry reconstruction of deformed clay analog models tends to improve at higher extension magnitudes because of fault and associated texture development on model surfaces. We anticipate that our results will help practitioners to improve the precision and reliability of photogrammetric data acquired in the analog modeling laboratory.