Capture overlapping views
Move around the subject and record photographs or video frames with steady coverage. Each view should share visual features with the next so the system can match them reliably.
3D capture explained
What is a realityscan, and how does it work?
What is a realityscan? It is a photogrammetry-based workflow that uses overlapping images of a real object or place to build a digital 3D representation. RealityScan helps turn ordinary capture into geometry, texture, and a model you can inspect or export.
- 2D
- Photos or video frames as input
- 3D
- Model generated from the capture
- Photogrammetry
- The reconstruction method
Related RealityScan topics
Three mechanisms behind the result
RealityScan is easier to understand as three connected operations: capture enough visual evidence, solve the camera positions, then build and refine the surface.
Solve position and structure
The reconstruction engine compares repeated details across images to estimate camera positions and form a sparse-to-dense representation of the subject.
Build the usable model
Matched information becomes a mesh with optional texture detail. You can then inspect the result, clean weak areas, and send the model into another 3D workflow.
The workflow in four steps
A successful RealityScan session depends more on consistent coverage than on a single perfect image.
- Choose a subject with visible detail
- 01
- Capture overlapping angles
- 02
- Review alignment and coverage
- 03
- Export or refine the model
- 04
From visual evidence to digital form
The change is not a flat image becoming magically accurate geometry. It is a series of visual observations being combined into a spatial model.
Captured views
More overlap generally gives the solver more evidence to compare, while hidden or reflective areas remain difficult.
Where the workflow is useful
The same capture principle serves different goals, but each use case has its own standards for cleanup, scale, and accuracy.
Maker or hobbyist
Capture a small object, prop, or sculpted form from several angles.
Create a reference model, shareable asset, or starting point for fabrication. The guide to realityscan for 3d printing covers the extra preparation needed before a physical result.
realityscan for 3d printingMobile creator
Use a phone-based capture session to document an object, location, or creative subject.
Turn a quick capture into a model that can be reviewed, edited, or passed to another application. A realityscan online free workflow may suit simple browser-based processing needs.
realityscan online freeTechnical artist
Generate rough digital references for environments, props, or early asset development.
Reduce manual blockout time while keeping expectations realistic about topology, texture quality, and hidden surfaces. RealityScan examples show the range of subjects a scan can represent.
realityscan examplesField researcher
Document a physical subject or setting when a conventional measuring workflow is slow or impractical.
Preserve a visual record that can support inspection and discussion, provided the capture has adequate coverage and the required scale is validated.
realityscan tutorialWhat RealityScan can and cannot infer
A scan is an interpretation of captured evidence, not a direct measurement of every part of an object. This distinction matters when deciding whether the result is suitable for reference, design, or production.
RealityScan is strong at
RealityScan needs care with
1
Visible textured surfaces
RealityScan is strong at
Repeated visual detail gives the solver useful points to match.
RealityScan needs care with
Plain, glossy, transparent, or repetitive surfaces can provide weak matches.
2
Overall shape
RealityScan is strong at
Many overlapping views can produce a convincing broad form.
RealityScan needs care with
Thin edges, deep cavities, and occluded areas may be incomplete.
3
Fast documentation
RealityScan is strong at
A capture can preserve a subject without modeling every feature by hand.
RealityScan needs care with
A quick capture does not automatically provide engineering-grade dimensions.
4
Texture reference
RealityScan is strong at
Images can contribute appearance information alongside geometry.
RealityScan needs care with
Uneven lighting, shadows, and motion can create visible texture defects.
5
Video as source material
RealityScan is strong at
Video can provide many frames when movement is controlled and coverage is consistent.
RealityScan needs care with
Blur, changing exposure, and unstable footage can reduce alignment quality.
6
Creative and draft assets
RealityScan is strong at
The result can be a useful base for visualization, blocking, and reference.
RealityScan needs care with
Production use may still require retopology, cleanup, scaling, and texture work.
7
Open-ended subjects
RealityScan is strong at
Objects, sculptures, and environments can all be approached with the same core method.
RealityScan needs care with
The best capture path depends on size, lighting, access, and the intended output.
Frequently asked questions
RealityScan is a photogrammetry workflow for creating a 3D representation from overlapping photographs or video frames. It estimates the position of each view, finds matching visual features, and uses that evidence to build geometry and appearance information.
Not exactly. A 3D photo usually suggests a single viewpoint or depth effect, while RealityScan combines multiple views to reconstruct surfaces from different angles. The result can be a mesh or point-based representation that is more useful in a broader 3D workflow.
You need a subject with enough visible detail, a series of overlapping images or suitable video frames, and consistent coverage around the subject. Good lighting, controlled movement, and attention to hidden areas usually matter more than simply taking a larger number of images.
A scan can support visual reference, creative blockouts, documentation, environment work, and some 3D printing workflows after cleanup and scale checks. It should not be treated as automatically accurate for every measurement or production requirement.