📸 Photogrammetry
From photos to 3D models — fundamentals and techniques
What is Photogrammetry?
Photogrammetry is the technique of reconstructing precise 3D models from 2D photos. The software analyses overlapping images, finds common points and computes geometry and texture from them.
Applications: film/VFX, gaming, archaeology, architecture, surveying, e-commerce, museums.
Capture Fundamentals
Camera Settings
- Aperture: f/8 to f/11 for maximum depth of field
- ISO: As low as possible (100–400)
- Focus: Manual, locked on the object
- Format: RAW if possible, otherwise the highest JPEG quality
Lighting
Ideal: Diffuse, even light without hard shadows.
- Overcast sky (outdoors) = perfect
- Softboxes or bounces (studio)
- Avoid direct sun → hard shadows
💡 Pro tip: For glossy objects, use cross-polarization: a polarizing filter on the camera plus one on the light eliminates reflections.
Overlap
The images must overlap heavily so the software can match points between them.
- Minimum: 60% overlap
- Optimal: 70–80% overlap
- Rule of thumb: one photo every 10–15°
Capture Methods
- Turntable: the object rotates, the camera stays fixed. Good for small objects.
- Walk-around: you move around the object. Good for large objects/scenes.
- Drone: for buildings, landscapes and large structures.
Software Workflow (Overview)
- Import: load the photos into the software
- Alignment: compute the camera positions
- Dense Cloud: generate the point cloud
- Mesh: build a surface from the points
- Texture: project the photos onto the mesh
- Export: OBJ, FBX or GLTF
📖 Step by step: The detailed Photogrammetry Tutorial walks you through a complete run with specific clicks and settings. For an overview of both techniques, see the Workflow Guide.
Common Problems
- Holes in the mesh: more photos from different angles
- Poor texture: improve the lighting, avoid shadows
- Alignment fails: more overlap, better contrast
- Glossy surfaces: polarization or scanning spray