# Photogrammetry

Photogrammetry is a technique that measures and reconstructs real-world objects from photographs. Software finds the same features in many overlapping photos, calculates the camera positions, and triangulates millions of 3D points, which it turns into a textured mesh. It needs dozens to hundreds of images of the real subject.

_AI Generation · updated 2026-11-16_

---

## What is Photogrammetry?

**Photogrammetry** is the science of measuring from photographs. In 3D work, it means rebuilding a real object, person, or location as a textured 3D model from a set of overlapping photos. It is the technique behind many scanned game assets, film environments, and digital copies of heritage sites.

Photogrammetry measures what is really there. If the photos are good, the result has accurate proportions and photographic surface detail.

## How Photogrammetry Works

1. **Capture**: photograph the subject from many angles with strong overlap and even, diffuse light.
2. **Feature matching**: the software finds the same visual features in different images.
3. **Camera solving**: it calculates where each photo was taken (structure from motion).
4. **Dense reconstruction**: it triangulates millions of points and builds a dense [mesh](/glossary/mesh).
5. **Texturing**: it projects the photos onto the mesh to create the color texture.
6. **Cleanup**: artists remove the background, fill holes, [remesh](/glossary/remesh) or [retopologize](/glossary/retopology), and bake the detail into maps.

## Photogrammetry vs AI Image-to-3D

| | Photogrammetry | AI [image to 3D](/glossary/image-to-3d) |
|---|----------------|-----------------|
| Input | Dozens to hundreds of photos of the real subject | One image, or a few views |
| Subject | Must physically exist | Photo, drawing, render, or concept art |
| Hidden sides | Must be photographed | Predicted by the AI |
| Accuracy | Measured, high | Plausible, estimated |
| Output | Very dense scan that needs cleanup | Mesh at a chosen face count, with UVs and textures |

Photogrammetry is the right choice when you need a measured digital copy of a real thing. AI reconstruction is faster when you need a usable asset from limited references. [Multi-view 3D](/glossary/multi-view-3d) sits between the two: a few photos instead of one, with AI filling in the gaps.

## Photogrammetry and Sloyd

Sloyd does not run photogrammetry. [Sloyd Image to 3D](/image-to-3d) uses AI to build a textured model from a single photo, drawing, or sketch, and multi-image input lets you add up to three extra views for better coverage of the shape. Because Sloyd builds the model at the face count and topology you choose, you skip the cleanup that raw scans need. If you need exact measurements of a physical object, a photogrammetry or scanning tool is still the better choice.

## FAQ

### What is photogrammetry?

Photogrammetry is the process of creating 3D models from photographs. You photograph a real object or place from many angles, and software matches the images to calculate the shape and color of the surface. It is used for game assets, film sets, heritage scanning, surveying, and product visualization.

### How many photos do I need for photogrammetry?

Typically between 30 and several hundred, depending on the size and complexity of the subject. Each part of the surface should appear in several photos with plenty of overlap between neighboring shots, usually around 60 to 80 percent.

### What objects are hard to capture with photogrammetry?

Shiny, transparent, very dark, or plain surfaces without texture are difficult, because the software cannot find features to match. Moving subjects, thin parts such as hair or wires, and changing light also cause holes or noise in the model.

### Is AI image-to-3D the same as photogrammetry?

No. Photogrammetry measures the real geometry from many photos of a physical subject. AI image-to-3D predicts a plausible 3D shape from one or a few images, including the sides the camera did not see. It is much faster and needs no capture session, but it estimates rather than measures.

### Do photogrammetry models need cleanup?

Usually. Raw scans are very dense, can have holes and baked-in lighting, and include parts of the background. Artists clean, retopologize or remesh, and re-bake the textures before using a scan in a game or real-time app.

## Related terms

- https://www.sloyd.ai/glossary/multi-view-3d
- https://www.sloyd.ai/glossary/image-to-3d
- https://www.sloyd.ai/glossary/retopology
- https://www.sloyd.ai/glossary/remesh

---

Source: https://www.sloyd.ai/glossary/photogrammetry
