Photogrammetry

Photogrammetry is the technique of reconstructing 3D geometry and texture from many overlapping photographs of the same subject. It's accurate for capturing existing buildings, sites, and objects, but requires dozens to hundreds of photos. AI image-to-3D, by contrast, infers a plausible 3D model from a single render or photo — orders of magnitude faster, with the trade-off that it's a generative inference rather than a metric capture.

In context

Architectural visualization uses both. Photogrammetry captures site context — terrain, neighbouring buildings, existing structure — for grounding new designs. AI image-to-3D (Renovato's mode) generates volumes from a hero render so designers can iterate in 3D before commiting to a BIM model.

In Renovato

Renovato's 3D mode is AI image-to-3D, not photogrammetry. The atlas can ingest a photogrammetric site model as a separate node and composite the AI-generated building volume into it for context renders.