Combo scanning
Combo scanning uses complementary visual and laser observations of the same object within one coordinated reconstruction workflow.
ASCAND’s Vision path and Laser path do not produce the same kind of evidence. Vision can derive object volume and surface coverage from images, silhouettes and related visual observations. Laser triangulation derives surface samples where the projected laser line reaches the object and the camera can detect it under the modeled geometry.
A Combo workflow obtains related laser-enabled and laser-disabled observations, processes the evidence through its appropriate reconstruction paths and evaluates how the resulting geometry agrees, differs or fills a local gap. Only then can supported contributions be reconciled into a combined representation.
That distinction matters. Combo is not simply two point clouds placed in one file. Nor does the word “combined” guarantee that every region becomes complete or more accurate. If both methods miss a hidden surface, neither can supply it. If one observation is unreliable, combining it does not automatically make it true.
Compare Vision, Laser and Combo
The two paths contribute different kinds of geometric evidence
Every reconstruction method interprets observations under its own assumptions.
The Vision path can use object boundaries and image evidence gathered around the rotation. In a silhouette-based reconstruction, each view constrains the volume the object may occupy. Combining those constraints produces a visual hull: the largest volume consistent with the observed silhouettes. This can provide broad, coherent shape even where a surface has little texture, but it cannot reveal a concavity that never changes an observed outline.
The Laser path projects a plane of light onto the object. A detected position along the observed laser trace defines a camera ray. Intersecting that ray with the modeled laser plane provides a candidate surface sample. Repeating this around the rotation builds a laser-derived point cloud. This can contribute direct local surface evidence, but only where the laser reaches the surface, the camera sees the response and processing identifies it reliably.
The evidence sources therefore have different strengths and failure modes:
| Evidence source | What it can contribute | Characteristic boundary |
|---|---|---|
| Vision and silhouette evidence | Coherent object volume, broad shape and visual coverage | Hidden concavities and segmentation errors can remain |
| Laser evidence | Local triangulated surface samples and strongly supported boundaries | Laser shadow, camera occlusion and material response limit coverage |
| Agreement between both | Additional support for a shared geometric interpretation | Agreement does not certify dimensional accuracy |
| Support from one path only | Potential geometry where the other path is incomplete | The contributing observation still requires evaluation |
Neither path is universally authoritative. A laser point is not dependable merely because it came from a laser, and a closed Vision model is not necessarily geometrically correct merely because its surface is continuous. Combo processing must consider the origin and quality of the evidence.
Assess whether an object suits the intended method
Related capture conditions make the evidence comparable
A Combo scan requires related observations of the same object with and without the laser contribution.
The steady camera, rotating object and coded turntable provide a controlled acquisition relationship. At a given part of the rotation, a laser-disabled observation records the object and scene without the projected line. A related laser-enabled observation records the scene with the line present.
That relationship can support several responsibilities:
- associating observations with a common rotational context;
- distinguishing a projected laser response from stable image content;
- keeping the object in a comparable pose and coordinate relationship;
- supporting separate Vision and Laser interpretation;
- making later geometric comparison more meaningful.
The laser-off observation is not merely an attractive color image, and the laser-on observation is not a finished depth map. They are inputs whose differences and shared context help processing identify the relevant evidence.
This also explains why capture stability matters. If the camera moves, the object shifts, the turntable relationship changes or the relevant code becomes unreadable, the observations may no longer describe the same controlled state closely enough for the intended comparison. Illumination, focus, framing, exposure, object stability and laser alignment can all affect what remains usable.
The conceptual relationship does not define current operating steps or safety conditions. Hardware setup, laser use and capture must follow the verified instructions for the installed equipment and current platform.
Follow the verified Laser and Combo workflow
Vision and Laser are reconstructed before they are fused
ASCAND’s multi-evidence architecture deliberately keeps reconstruction sources separate long enough for them to be assessed.
Vision observations first produce Vision-derived geometry under the selected reconstruction path. Laser observations first undergo laser-profile identification, triangulation, transformation and point-cloud preparation. Each result retains meaning because its origin remains identifiable.
Delayed combination serves an important purpose. If all observations were mixed immediately, a clean result could conceal which method supported a surface, which method disagreed and where an artifact entered the pipeline. Independent outputs allow comparison, quality evaluation and source-aware decisions before a unified representation is formed.
Conceptually, the architecture separates:
- captured image observations;
- orientation interpretation;
- Vision reconstruction;
- Laser reconstruction;
- point-cloud or geometry preparation;
- evidence comparison;
- reconciliation and later surface generation.
The branches are related, but one is not simply a filter applied to the other.
A Laser workflow may, for example, use laser-on and laser-off observations to isolate the projected response without completing a separate Vision reconstruction. The absence of a Vision-derived output would not by itself mean that Laser processing failed. A true combined result, however, requires usable geometry from the relevant independent sources before their contributions can be reconciled.
Review the documented processing paths
From paired observations to a combined representation
A public conceptual model of Combo processing contains seven stages.
1. Acquire related observations
The object is recorded within the controlled camera–turntable relationship, including the observations required for the selected Laser and Vision processing paths.
2. Interpret shared rotational context
The captured sequence and visible coded reference provide orientation-related evidence. This helps associate observations and reconstructed samples with a common object relationship.
3. Reconstruct Vision evidence
The Vision path interprets applicable image, boundary and silhouette observations to create its independent geometric representation.
4. Reconstruct Laser evidence
The Laser path identifies usable projected-line observations, triangulates candidate surface samples and accumulates them in the relevant coordinate system.
5. Prepare and compare geometry
The independent representations are cleaned or normalized as required and evaluated for alignment, local support, overlap, gaps, outliers and disagreement.
6. Reconcile supported contributions
Combination logic determines where evidence agrees, where one path can contribute useful support and where a contradiction requires rejection or further evaluation. This is evidence reconciliation, not indiscriminate averaging.
7. Generate and inspect the combined output
The reconciled geometry can support a combined representation and later surface reconstruction. The result still requires inspection against the object, the independent outputs and the intended downstream use.
These stages explain responsibilities, not proprietary thresholds, confidence weights, interface controls or processing-time promises. Available output variants and current workflow details remain platform- and version-dependent.
Understand output families and post-processing
Agreement increases confidence; disagreement requires evaluation
When independent methods support approximately the same local surface, their agreement provides stronger evidence for that interpretation. It does not, however, turn the result into a certified measurement.
Three broad situations can occur.
Both paths support a region. Their geometry may agree closely enough to reinforce a shared surface or boundary. Remaining systematic errors can still affect both.
One path supports a region more strongly. Vision geometry may provide continuity where usable laser samples are absent. Laser geometry may supply local surface or boundary evidence where a Vision hull is too broad. The contributing source must still be evaluated on its own terms.
The paths conflict. Surfaces may be offset, differently shaped or supported by incompatible observations. Possible causes include alignment error, segmentation error, weak laser detection, reflections, object movement, uncertain orientation, outliers or a limitation of one reconstruction model.
Conflict does not automatically identify a winner. A source-aware workflow should consider acquisition origin, local consistency and supporting observations before accepting, adjusting or rejecting geometry.
For the same reason, a combined output should not be described as a simple union. Concatenating all points would retain duplicates, incompatible surfaces and noise. Simple averaging could move a well-supported boundary toward a weak or erroneous estimate. Reconciliation must preserve useful evidence while managing redundancy and contradiction.
Diagnose common Laser and Combo problems
Complementary evidence can improve coverage without removing uncertainty
Combo scanning is useful because the methods do not fail in exactly the same way.
Where suitable evidence exists, a combined workflow may:
- preserve well-supported laser-derived surface regions or boundaries;
- use Vision-derived geometry where laser coverage is absent or incomplete;
- reduce gaps that affect only one reconstruction path;
- compare independent interpretations of the same object;
- expose both independent and combined results for evaluation;
- provide more coherent input for later surface generation.
These are possible contributions, not universal outcomes.
Some limitations are shared. A surface hidden from both the camera and laser remains unobserved. A deep region that never affects a silhouette and never receives a visible laser line cannot be recovered as measured geometry. Transparent, translucent, highly reflective, very dark or otherwise optically difficult surfaces may disrupt both paths in different ways.
Other problems can spread through combination. Incorrect camera or laser geometry can misplace points systematically. Object movement can weaken correspondence between observations. Poor segmentation can distort the Vision volume. False or broadened laser responses can introduce outliers. Misaligned independent reconstructions can create seams or doubled surfaces.
Later meshing, smoothing and hole repair may make a result look continuous, but those operations do not add missing observations. A visually closed surface can still contain inferred or incorrect geometry.
Judge Combo by the intended use and by evidence visible in the actual result—not by the scan-type label alone.
Choose an evidence method before choosing a workflow
Vision, Laser and Combo are different evidence strategies.
Choose Vision when the documented image and silhouette path fits the object and intended result. Choose Laser when projected-line observations can supply the surface evidence you need. Choose Combo when both paths can contribute meaningful, comparable evidence and the value of reconciliation justifies the additional acquisition and evaluation.
Do not confuse Combo with Multi-Scan or Multi Merge. Combo concerns different evidence types associated with a related object capture. Multi-Scan concerns separate reconstructions—often with the object placed in different orientations—to reveal surfaces that one setup could not see. Those reconstructions require sufficient overlap and their own alignment and merge logic.
For method selection, continue with the comparison article. For additional object orientations, continue with Multi-Scan. For operation, use the current setup and capture tutorials.
Compare Vision, Laser and Combo
Understand Multi-Scan and Multi MergeLaser triangulation reconstructs visible surface points from a controlled geometric relationship between a camera and a projected laser line.