← Applied Spatial Intelligence Labs

South House — a building, rebuilt in 3D from photographs

No scanner, no LiDAR, no drone. Just 128 ordinary photos walked around the structure. Our spatial pipeline recovers where every camera stood and reconstructs the scene as a 61,000-point colored 3D model you can orbit right here in your browser — the same photogrammetry core behind ASI Labs digital twins.

Reconstruction · points Input · photos Runs · in-browser WebGL
— pts — fps
Point size
Drag to orbit · scroll to zoom · right-drag to pan
Decoding reconstruction…

Every colored dot is a real 3D point triangulated from two or more photographs. Positions are quantized to keep the page light — the full-resolution model carries millimeter-consistent geometry and feeds a dense Gaussian-splat or textured-mesh twin.

From a handful of photos to a measurable model

The same four-stage pipeline runs on a house, a facade, a field structure, or an industrial site. Photographs go in; a geometry you can measure, navigate, and simulate against comes out.

Step 1

Capture

128 overlapping handheld photos around the building. No specialized rig — a phone is enough.

Step 2

Solve cameras

Feature matching + structure-from-motion recover the exact 3D pose of every shot.

Step 3

Triangulate

Shared features are lifted into a colored 3D point cloud — the model you are orbiting.

Step 4

Twin

Densify into a Gaussian-splat or textured mesh for walkthroughs, measurement, and simulation.

Reconstructed 3D points
Source photographs
3072×2304
Pixels per photo
100%
Camera poses recovered

The input — ordinary photographs

A sample of the source frames. Nothing here is a render: each is a real photograph the pipeline matched against its neighbors to place in 3D space.