Two scans were performed: a Lidar scan (for a more accurate-to-reality digital model) and a path-based photogrammetry scan. These registrations were superimposed to understand what new digital-only geometries were being produced by the photogrammetry scan’s misregistrations. Regions were selected for further architectural articulation based on: locating areas of extreme difference between the photogrammetry and Lidar model, and the region’s ability to host an architectural intervention consideringfactors such as safety, accessibility, existing program, and circulation.

The array of points produced by the photogrammetry scan resulted from a broad range of conflicting environmental and software factors, mainly pertaining to the mechanics of how the specific photogrammetry software processes digital photographs to produce a digital model. Photogrammetry is dependent on: light and shadow to determine surfaces (from its perception of texture), form, position of perceived objects, relativity, and image metadata (EXIF) such as time taken, aperture, and other camera-based settings. As a result, items such as “wavy” or “undulating” surfaces appear as the software attempts to consolidate its understanding that there is a large continuous surface at that given location but to generate the surface completely requires a degree of differentiation between light and shadow to determine its precise boundaries. Many of the other types of phenomena such as the appearance of “smoky” objects were hypothesized to result from variances in lighting between images of the same object and camera misplacement caused by incorrect correspondence between objects that are similar in appearance but occur at different real-world locations.