Photogrammetry has become the default real-world capture technique for web 3D and AR, mostly because the hardware barrier disappeared. Any modern smartphone can capture an object well enough; everything else happens in software. The workflow is consistent across tools: 50 to 300 overlapping photos, the software derives camera positions through Structure from Motion, builds a dense point cloud, meshes it, and projects the original photos back onto the surface as textures.
The tooling now spans every price point. RealityCapture from Epic Games is free for most use cases under a pay-on-export model, Agisoft Metashape is the professional staple, and mobile apps like Polycam, Luma AI, KIRI Engine and Scaniverse handle smaller objects directly on the phone. Apple’s Object Capture API, available in Reality Composer’s iOS app and as a RealityKit API on macOS, uses the LiDAR Scanner on Pro devices to refine the result. Photogrammetry meshes almost always need cleanup in Blender (decimation, retopology, texture baking) before they ship to the web as GLB and USDZ.
A real example: Drone e-motion uses drone-based photogrammetry to capture historic buildings, including the windmill “Moulin a vent de Frouville”, and publishes them as Web AR experiences through PausAR Viewer on WordPress. We documented the full workflow: how Drone e-motion uses PausAR for 3D and AR experiences from drone photogrammetry.
| Property | Photogrammetry | LiDAR scanning |
|---|---|---|
| Hardware | Any camera or smartphone | LiDAR sensor (Pro iPhones, iPad Pro, dedicated scanners) |
| Strength | Realistic textures from the original photos | Accurate measurements and instant depth |
| Weakness | Struggles with reflective or transparent surfaces | Less texture detail, mostly spatial structure |
| Capture time | Slower (many photos plus processing) | Near-instant |
| Best for | Visual fidelity, products, heritage buildings | Spatial mapping, room scanning, AR placement |
Both capture the real world, but differently. Photogrammetry derives 3D from many overlapping photos and carries rich textures from the original images. LiDAR measures distance directly with light pulses, which is fast and accurate for spatial structure but carries less visual detail. Many modern workflows combine both, like Apple Object Capture, which uses the LiDAR Scanner to refine a photogrammetry result.
Yes. Apps like Polycam, Luma AI, KIRI Engine and Apple Object Capture (in the Reality Composer iOS app) produce usable results from phone photos. Quality depends on lighting, surface type and how many angles you cover. For high-end professional captures, desktop tools like RealityCapture or Metashape give more control over the mesh and textures.
Clean up the mesh in Blender, then export to GLB and USDZ. Upload both files to the PausAR Viewer Elementor widget, and the model becomes an interactive 3D viewer with optional Web AR. Drone e-motion is a real example of exactly this workflow, capturing historic buildings via drone photogrammetry and publishing them on WordPress.
You are currently viewing a placeholder content from YouTube. To access the actual content, click the button below. Please note that doing so will share data with third-party providers.
More InformationYou need to load content from reCAPTCHA to submit the form. Please note that doing so will share data with third-party providers.
More Information