The 6-Step Workflow Behind Automated 3D Scan-to-Measurement — And Why It Actually Works
3DRevopointLet's start with a scenario that's probably familiar.
A batch of parts comes off the line. You walk over to the CMM station and the queue is already backed up — two parts waiting, one on the table, the operator mid-program. You do the mental math. Results won't be ready until mid-afternoon, and the next batch is already running.
If there's a deviation in this batch caused by thermal stress or mold wear deformation, you're not going to find out in time to do anything about it.
This isn't a workflow problem you can solve by hiring another CMM operator. It's a structural mismatch between how CMMs work and what modern production lines actually demand. Metrohub was built specifically around that gap — and the six-step workflow it runs is worth understanding in detail, because the logic behind each step matters as much as the step itself.
First, a Word on Why CMM Alone Isn't Enough Anymore
To be clear: CMMs are precise, repeatable instruments. Nobody's disputing that.
The limitation isn't accuracy — it's architecture. A CMM measures at discrete touch-points you specify in advance. You get data on the 32 features you programmed. What's happening on the rest of the surface is invisible. For a simple prismatic part with well-defined critical features, that's fine. For a casting with complex freeform geometry, or a stamped panel where springback can show up anywhere, you're flying partially blind.
There's also the throughput problem. CMMs are serial by design — one machine, one operator, one part at a time, with non-trivial setup time every time the part number changes. For high-mix production environments running 10 to 20 different part numbers per shift, that model doesn't scale. The inspection station becomes the bottleneck, and defects travel downstream before the data catches up.
Metrohub's approach is different: full-surface, non-contact, robotic inspection that runs continuously between parts without operator involvement. Here's how it works.
Step 1 — Setup: One-Time Configuration, Indefinite Reuse
Before the first scan, an operator uses Metrohub's interface to manually teach the scanning trajectory, configure scanner parameters — working distance, resolution, exposure — and teach the robot its positioning sequence.
For most production lines running recurring part numbers, this is a one-time investment. The program is saved and recalled for every subsequent run. No fixture to rebuild, no probe stylus to swap, no re-programming from scratch when the same part comes back next week.
This is a meaningful difference from CMM setup, where each new part or fixture configuration requires hands-on metrologist time. The robot adapts to the part. The part doesn't need to adapt to anything.
Step 2 — 3D Scan: Automated Data Acquisition via Track Bot
Once the program is loaded, Track Bot — the automated scanning software — takes over completely.
The robotic arm moves through a manually taught multi-angle path. Before production scanning begins, an operator teaches the robot its trajectory — positioning the arm at each required angle to ensure deep pockets, undercuts, internal chamfers, and other features that a fixed scanner would miss are all covered. Once the teaching is complete, the path is saved and repeated automatically for every subsequent part.
One technical detail worth highlighting: Metrohub uses blue laser technology. This matters on a production floor because real parts aren't always cooperative surfaces. Cast alloys, anodized aluminum, and black rubber seals are all materials that are either too dark or too reflective for conventional structured-light or white-light scanners. Blue laser handles most of these without anti-reflective spray. In a production environment where stopping to prep surfaces adds minutes per part, that's not a minor convenience — it compounds across an entire shift.
The output is a dense point cloud covering the complete part geometry. Not a sampled set of coordinates — the full surface.
Step 3 — Post-Processing: From Raw Points to a Usable Mesh
Raw point cloud data isn't ready for measurement. It contains noise from ambient vibration, minor reflections, and registration errors between scan frames. Track Bot automatically runs point cloud registration and fusion, followed by mesh generation, before the data moves downstream.
There's a subtlety here that's easy to underestimate: uneven point cloud density.
Near deep cavities, density tends to drop off. If you apply uniform downsampling during meshing — which is the default behavior in a lot of processing pipelines — those low-density regions produce false surface bumps that show up as deviations in the color map. You end up chasing a "defect" that's an artifact of the data processing, not the part.
Track Bot offers two downsampling methods. Uniform downsampling reduces point count evenly across the entire surface — straightforward, but it treats a flat face and a complex fillet the same way. Geometric downsampling is the more precise option: point density is preserved along edges, fillets, and surface transitions where the geometry changes rapidly, and reduced in flat, featureless regions where additional points contribute no meaningful geometric information. The result is a clean, continuous mesh that accurately represents the part geometry without inflating file size.
Step 4 — Export to Revo Measure: No Handoff Friction
The mesh transfers directly into Revo Measure, Revopoint's metrology software, in a single click.
No STL export, no manual coordinate system re-alignment, no file format conversion, no renaming conventions to manage.
If you've worked in a mixed-vendor workflow — scanning software from one company, measurement software from another — you know exactly why this matters. The handoff between them typically involves exporting STL, importing into GOM or PolyWorks, and manually re-defining the coordinate system from scratch. On a good day, that's 15 minutes. On a bad day, it's 30, plus the risk of subtle alignment errors that are easy to miss and hard to trace.
One-click transfer eliminates those intermediate steps entirely. That said, the coordinate system preserved at this stage is scanner-based. Dimensional measurement can proceed directly on that basis, or you can define a custom coordinate system inside Revo Measure to align with your engineering drawing datums — whichever works for your inspection workflow. That step happens inside Revo Measure, but at least you're starting from clean, unmodified scan data rather than a file that's been exported, re-imported, and re-aligned by hand.
Step 5 — Measurement Analysis: Where the Data Becomes Actionable
Inside Revo Measure, the imported point cloud or mesh is aligned to the original CAD model using Best-Fit Alignment. From there, the analysis options cover the full range of what dimensional inspection requires.
Full-surface color deviation maps are the most immediate output. The entire part surface is color-coded against nominal — red for positive deviation, blue for negative, green for in-tolerance. This is qualitatively different from a CMM report in a way that matters: a list of 32 point measurements tells you whether those 32 points are in tolerance. A full-surface deviation map tells you the shape of the problem.
A part affected by thermal stress shows consistent blue zones on the outer faces — you can read the contraction pattern directly from the color map. A part with mold wear deformation shows deviation that concentrates predictably along the affected surfaces. These are process signatures, and they're visible at a glance in a way that a table of point measurements cannot replicate.
Beyond the color map:
● GD&T callout evaluation — flatness, cylindricity, true position, and profile of a surface evaluated against engineering drawing tolerances
● Cross-sectional analysis — bore geometry on machined housings, wall relationships, and any feature that requires a sectional view to evaluate properly
● Feature-level dimensional measurements — distance, angle, bore diameter, and other nominal dimensions
The color map is also your earliest warning system for process drift. By the time a CMM report has accumulated enough data points to flag a trend, you've often already produced scrap. A full-surface scan catches the deviation pattern earlier, when it's still a process adjustment rather than a quality event.
Step 6 — Report Generation: Consistent, Automatic, Audit-Ready
The report contains full-color deviation maps with scale bar and tolerance limits, individual GD&T feature results with nominal, actual, and deviation values, cross-sectional profiles at specified planes, and a summary PASS/FAIL table.
The default format is consistent across every part, every operator, every shift — and for teams that need something different, the report template is fully customizable. Company logo, measurement items, layout, output language — these can be configured to match internal documentation standards or customer-specific requirements.
That consistency matters more than it might seem. For ISO or IATF audit trails, consistent documentation is a requirement, not a preference. For customer-facing inspection reports, a structured PDF that looks the same every time is far more useful than a raw data file that requires interpretation.
No manual formatting. No copy-pasting values into a template. The report is generated and ready.
How This Plays Out in Real Production
Forging and casting inspection: After cleaning, parts go directly to the Metrohub station. The robot scans the full part in minutes. Revo Measure checks key dimensions — center distances, bore allowances, and critical surface profiles — and flags surface defects like cold shuts or underfill. Critically, the scan data can feed back into machining parameters: if a casting is consistently running 0.3 mm thin on one face, the CNC offset gets adjusted before the batch hits the mill. You're correcting upstream, not discovering downstream.
Automotive sheet metal: Door panels, body-in-white components, structural stampings — complex freeform surfaces that a CMM can only sample. The Metrohub scan captures the full panel geometry. Revo Measure compares it against the Class A surface definition and surfaces any springback or oil-canning that would affect fit or appearance. You see the whole panel, not 40 points on it.
Precision machined components: Mold inserts, engine blocks, gearbox housings — parts that deviate from design due to thermal growth during machining. The system supports 100% inspection, every part, without slowing the line. Not a statistical sample. Every part.
A Few Questions Worth Addressing Directly
Does automated 3D scanning replace CMM entirely?
For most dimensional inspection on complex parts, yes — and it provides more information. The honest exception: applications requiring sub-micron accuracy on simple geometries (gauge blocks, precision bores) still belong on a CMM. For everything else, full-surface scanning is faster, more comprehensive, and doesn't require a dedicated metrologist to operate.
What about challenging surfaces — mirror-polished chrome, jet-black rubber?
Blue laser handles the majority of production surfaces without prep. True edge cases may still need a light application of scanning spray, but that's the exception. Most parts that come off a real production line don't require it.
How fast is the full cycle?
For a typical machined part in the 200–500 mm range: scan start to PDF report in under 10 minutes. A comparable CMM inspection of the same part — 40 to 50 touch points — typically runs 25 to 45 minutes, not counting setup. The difference compounds quickly at production volumes.
The Bottom Line
The Metrohub workflow doesn't replace engineering judgment. It gets better data to the engineer faster, so that judgment can be applied before a deviation becomes a production event.
The system handles the repetitive work: moving the scanner, capturing geometry, aligning to CAD, generating the report. The engineer decides what the data means and what to do about it. That division of labor is the point.
If your inspection process is a consistent bottleneck — if defects are showing up at assembly instead of at the inspection station, if your CMM queue is always backed up, if changeover time between part numbers is eating into your shift — this is the problem the system is built to solve.
Reach out to the Revopoint-Robot technical team to discuss your specific part types and production volumes. The right configuration depends on your part envelope, surface characteristics, and cycle time requirements.