ASCAND and photogrammetry
ASCAND is not accurately described as a conventional photogrammetry system.
The two approaches share an important foundation: both can use camera observations of an object from multiple directions to help reconstruct three-dimensional geometry. But they establish and interpret those views differently.
In a conventional photogrammetry workflow, the camera usually moves around a stationary subject—or many photographs are taken from changing positions. The software identifies visual features that appear in overlapping images, determines which features correspond and estimates the camera position associated with each image.
ASCAND structures acquisition differently. The smartphone remains steady while the object rotates on a coded turntable. The recorded video contains changing views of the object together with a visible orientation reference. The physical capture relationship is therefore deliberately constrained rather than left entirely for the software to infer from an unconstrained set of photographs.
ASCAND can still use principles associated with photogrammetry and computer vision inside its wider reconstruction architecture. That does not make conventional photogrammetry the complete or most accurate category for the system. ASCAND also uses silhouette-informed volumetric reasoning and can incorporate optional laser-derived evidence.
The precise description is:
ASCAND is a structured, turntable-based optical 3D scanning workflow with multiple documented reconstruction methods.
Start with the broader definition of 3D scanning
What conventional photogrammetry must determine
Photogrammetry reconstructs spatial information from multiple overlapping images. Although implementations differ, a typical object-reconstruction workflow must solve several related problems.
Find recognizable image features
The software looks for visually distinctive regions—such as corners, texture changes, edges or local intensity patterns—that can be recognized in more than one image.
An image can contain millions of pixels without providing millions of reliable geometric observations. Uniform, reflective, transparent or repetitive surfaces may provide fewer stable correspondences than a textured, matte surface.
Match observations across images
The system determines which detected features are likely to show the same physical location. Correct correspondences constrain the geometry. Incorrect correspondences can distort the reconstruction or cause parts of it to fail.
Estimate the viewpoints
The software determines a spatial arrangement of camera positions and orientations that explains the confirmed correspondences. In a conventional moving-camera workflow, those viewpoints may not have been recorded independently; they are part of the reconstruction problem.
Reconstruct points and surfaces
Once the spatial relationship among observations is sufficiently constrained, the system can estimate three-dimensional points and increase the density of the representation. Later stages may clean the point cloud, build a mesh and add appearance information.
This is an inference process. It searches for geometry and viewpoints that are consistent with the image evidence. More photographs do not automatically solve low texture, motion blur, reflections, changing illumination, weak overlap or hidden surfaces.
Review the wider family of 3D scanning methods
ASCAND controls the acquisition relationship
ASCAND changes the physical arrangement before reconstruction begins.
The smartphone is held steadily in front of the turntable. The object is placed on the turntable and rotates while the phone records a video. As the object turns, the camera observes it from a changing relative direction even though the camera does not travel around it.
The coded band around the turntable provides visible information associated with rotational orientation. The known turntable geometry and captured reference information support a structured camera–turntable coordinate relationship.
This has several practical consequences:
- The user does not need to walk around the object while maintaining an improvised camera path.
- Views arrive as an ordered video sequence rather than an unrelated photo collection.
- The object rotates around a defined physical setup.
- The turntable and its coded reference contribute information beyond the object’s natural visual texture.
- The same basic acquisition relationship can support different documented reconstruction paths.
Controlled acquisition reduces some variables; it does not eliminate uncertainty. The phone still needs stable framing and usable focus. The complete required turntable area must remain visible. The object must remain stable relative to the platform. Lighting and surface behavior still affect the observations. Hidden geometry remains hidden unless another capture orientation exposes it.
The coded reference also should not be described as a magical replacement for reconstruction. Orientation information helps relate observations. Geometry must still be derived from the evidence available to the selected method.
Learn why the camera stays still
Understand structured rotation and Gray Code
The overlap is real—but it is not the whole architecture
The clearest comparison separates shared principles from different responsibilities.
| Question | Conventional object photogrammetry | ASCAND structured workflow |
|---|---|---|
| How are different views created? | The camera commonly moves around a stationary object. | The smartphone remains steady while the object rotates. |
| How is view orientation established? | Image correspondences commonly help estimate changing camera poses. | The controlled turntable relationship and coded rotational reference provide structured acquisition information; individual reconstruction paths may still perform geometric estimation. |
| What visual evidence may be used? | Overlapping image features and appearance consistency. | Depending on the documented path: image features, segmentation, silhouettes and other computer-vision evidence. |
| Is active optical evidence available? | Not in passive photogrammetry itself. | The optional Laser Extension can contribute projected-line evidence through laser triangulation. |
| Does one method define the complete system? | Photogrammetric reconstruction is the principal method. | ASCAND is designed around distinct evidence and reconstruction paths that may later be evaluated or combined. |
| Are all surfaces guaranteed? | No. Visibility, texture, overlap and surface behavior limit reconstruction. | No. Visibility, method choice, surface behavior, capture conditions and reconstruction limits still apply. |
This is why two apparently conflicting statements can both be true:
- ASCAND Vision processing can use photogrammetric and computer-vision principles.
- ASCAND as a complete system is not conventional photogrammetry.
The first statement describes techniques inside an architecture. The second describes the architecture and user workflow as a whole.
ASCAND does not rely on one reconstruction method
ASCAND is designed as a multi-evidence reconstruction framework. Its documented processing architecture distinguishes among sources rather than treating all geometry as if it were obtained in the same way.
Image-correspondence evidence
Photogrammetric or computer-vision processing can identify corresponding features across normalized observations, estimate consistent viewpoints and reconstruct surface points. This path depends on recognizable appearance across views.
Silhouette and volumetric evidence
Vision processing can also use segmentation and silhouettes to constrain occupied volume. Voxel carving begins with a volume around the object and removes regions inconsistent with the observed foreground and background across known orientations.
This visual-hull reasoning is fundamentally different from feature matching. It can constrain overall shape where reliable object outlines are available, but silhouettes do not reveal every concavity or hidden surface.
Laser-derived evidence
With the appropriate extension and documented workflow, a projected laser line provides active optical evidence. Its observed position, together with the calibrated geometry, supports triangulation of visible surface locations.
Later evaluation and combination
These paths should remain conceptually separate because they support geometry differently. Image correspondence infers points through agreement across views. Silhouettes constrain occupied space. Laser triangulation derives positions from an observed projected line.
Independent representations can be evaluated, processed and—where the selected workflow supports it—combined later. Agreement can strengthen confidence; disagreement can reveal uncertainty. Combining sources is not proof that every region is accurate, and not every scan type should be assumed to execute every reconstruction path.
Compare Vision, Laser and Combo evidence
Explore Vision scanning and voxel carving
What the difference means in practice
The structured workflow changes what the user controls, but it does not make object choice irrelevant.
You control a station, not a camera walk
The primary physical task is to establish a stable relationship among the smartphone, complete turntable and object. During capture, that relationship should remain steady while the platform provides the rotation.
The turntable is part of the evidence
It is not merely a display stand. Cropping the required coded area, obscuring it or allowing the setup to shift can remove information needed to interpret the capture.
Object behavior still matters
The object should remain stable during rotation. Flexible parts, movement or changing shape mean that observations may no longer describe one consistent geometry.
Each reconstruction method has different sensitivities
Feature-based reconstruction benefits from stable, recognizable visual detail. Silhouette-based reconstruction depends on clear object boundaries and cannot recover hidden concavities from outlines alone. Laser evidence depends on visibility of the projected line and on the surface’s optical response.
Reflective, transparent, translucent, very dark or otherwise optically difficult materials may challenge one or more paths. A structured capture setup does not repeal those physical limitations.
A single orientation cannot expose every surface
The rotating object provides views around its vertical orientation, but an underside or deeply occluded region may never become visible. A complementary physical orientation and a documented Multi-Scan or Multi Merge workflow may provide additional evidence where sufficient overlap supports alignment.
Assess your object before capture
Review capabilities and limitations
Four descriptions that need correction
Misconception: “ASCAND is simply turntable photogrammetry.”
Accurate distinction: The shared use of multiple camera views is only part of the system. ASCAND adds a controlled steady-camera relationship, coded rotation, silhouette-informed reconstruction and optional laser evidence within a broader architecture.
Misconception: “The coded turntable means image reconstruction is unnecessary.”
Accurate distinction: Rotational reference helps locate observations in a structured sequence. It does not by itself describe the object’s surface. Geometry still comes from the selected visual, volumetric or laser evidence.
Misconception: “A fixed camera sees only one viewpoint.”
Accurate distinction: The camera’s physical position remains fixed, but the rotating object presents changing relative views. Reconstruction can model that relationship as observations around the object.
Misconception: “Browser processing turns every video into a complete model.”
Accurate distinction: 3D-Scan.Online processes captured material using the available workflow. Results remain limited by object suitability, visibility, capture quality, selected evidence, reconstruction behavior and downstream requirements.
ASCAND and 3D-Scan.Online also have separate responsibilities: Capture the object with ASCAND. Process the video on 3D-Scan.Online.
A visually plausible mesh is not proof of complete observation, certified dimensions, editable CAD or fitness for a particular use.
Understand point clouds and meshes
Review verified examples and their conditions
Continue with the evidence your project needs
Choose the next article by the question you need to answer:
- Why keep the smartphone fixed? Read Why the Camera Stays Still.
- What does the coded band contribute? Read Structured Rotation and Gray Code.
- How can silhouettes constrain shape? Read Vision Scanning and Voxel Carving.
- How does the projected line contribute geometry? Read Laser Triangulation.
- Which evidence should I choose? Use the Vision, Laser and Combo comparison.
- Will my object work? Begin with the Object Suitability assessment.
The useful question is not whether one method name can describe everything ASCAND does. It is which evidence a particular workflow obtains, how that evidence constrains geometry and whether the resulting representation supports the intended use.
Compare Vision, Laser and Combo3D scanning is a family of processes that uses observations of a physical object or environment to create a digital representation of its geometry.