TECHNIQUES USED TO BUILD PHOTOGRAMMETRIC MAPS/MODELS FOR ICHNOFOSSILS AT DINOSAUR STATE PARK, ROCKY HILL, CONNECTICUT

Photogrammetric and scanned data sets are increasingly used to measure, visualize, and replicate dinosaurian and other ichnofossils, both in the field and lab/museum. Here we detail our techniques for imaging and modelling a well-known Jurassic tracksite within the interpretive center at Dinosaur State Park (DSP). This is part of a multi-authored compendium of research on DSP currently under contract with Indiana University Press. The DSP tracksite is challenging to image/model due to dim and/or spot-lighting; as well, some 16% of the 335 m² tracksite is obstructed from view by a walkway, platform and diorama flooring. While our work is the first to digitally model/map this site, an important map by Galton and Farlow (2003) identifies and reports measures for many tracks on a map assembled from 1.5x1.5 m tracings ca. 1970. We rectify this trace map into our survey space and use it to assign track names and other attributes to many of the 751 Eubrontes tracks, 98 putative “swimming” tracks and several undescribed traces in our geodatabase. Moreover, our 0.75 billion-point model, which includes areas under the platform/walkway/diorama, is sufficiently dense (≈0.6 mm resolution) to segment out any track(s) for detailed topographic analysis. Our map/model is derived from ≈2300 full-frame images captured in 11 sub-areas and scaled using laser-surveyed ground control points (5-30/map area), selected locations from a 1.4x10⁶ point scan, and/or photogrammetric scale bars. Local accuracies vary from sub-mm for most of the tracksite to 1.5 cm for some areas obstructed by structural supports for the walkway. We describe several imaging/modeling/visualization techniques using examples of Eubrontes, “swimming”, and unidentified tracks. Dim/spot lighting and cramped imaging spaces are managed with masking, retexturing, and rescaling techniques. Cloud optimization approaches are summarized, as are methods to reframe segmented tracks and build statistical best-fit height maps suitable for rasterization and importing to ArcGIS for topographic and hypsometric analyses. We also touch briefly on animation techniques for digital models, all of which are equally applicable to larger mega-fossils and modern zoogeomorphology research.