A REFINED PHOTOGRAMMETRY WORKFLOW FOR CONSISTENTLY HIGH-RESOLUTION VIRTUAL 3D MODELS OF GEOLOGIC SPECIMENS SHARED ON SKETCHFAB

We present a novel double masking process in Agisoft Metashape that more accurately renders virtual 3D models of geologic specimens with large depth of field variances. The need for geologic models (hand samples, block models, etc.) recently peaked in association with the transition to distance learning formats necessitated by the COVID-19 pandemic. Few existing workflows, however, yield models that are of sufficiently high resolution to create 3D virtual versions of the block models used to foster an ability to visualize the complex interaction between deformed strata and the surface of the Earth. To address this need, we present a new photogrammetric workflow that maximizes lighting, photographic, and post-processing techniques for generating detailed models that fully capture large depth-of-field variances and reflective surfaces and make them serviceable even at the highest levels of magnification. These models have sub-millimeter detail and retain photographic focus on all surfaces with minimal textural smearing. In addition to an overview of this new workflow, we will share tips for (1) creating an inexpensive photo tent and turntable, (2) creating shadowless lighting, aligning camera positions, (3) using selective focus-picking and focus-stacking in image capture, (4) processing images in Adobe Lightroom, (5) batch-processing masking methods in Adobe Photoshop, (6) using focus-stacking for specimens with fine details and large depths of field, and (7) configuring a Metashape workflow that utilizes double masking. The double masking process first applies a background exclusion mask to images used in the model’s construction and then applies a secondary focus-selected mask to create a high-resolution photographic skin that appears focused from any viewing angle. We applied this technique to wooden models of deformed strata to create exercises that focus on developing 3D map visualization skills in remote teaching environments. Our models allow for quantification of the block topography, which enables us to ask quantitatively meaningful questions about the 3D structure. Not only can these virtual blocks help develop map reading skills in a remote learning environment, but the same exercises can provide a digital ramp to help prepare students to make the most of post pandemic field experiences.