Dimensional inspection
Gap, flush, geometry, measurement, tolerance verification.
Dimensional checks fail in production for reasons that have little to do with measurement accuracy: the part moves, the surface reflects, the geometry curves away from the camera, or the feature being measured is defined by the relationship between two edges rather than by a single edge.
Rule-based systems handle this by constraining the environment — fixturing the part, controlling the lighting, restricting the variants. That works until the line speeds up or the product changes.
FOCUS™ AI locates the features first and measures second, which keeps the measurement stable when the part presentation is not. Measurements are returned against your tolerance definitions and logged per part, so drift becomes visible before it becomes scrap.
What we inspect
- Gap and flushRelationships between adjacent panels, covers and trim.
- GeometryFeature positions, angles and profiles across the part.
- MeasurementDimensions taken at defined points, part after part.
- Tolerance verificationPass or fail against your own tolerance definitions.
How we get there
The first three steps of our deployment process. Nothing is promised before feasibility has been assessed.
- 01
Show us the challenge
Share your products, defect types, production environment and inspection requirements. Physical samples, images or video — whatever best represents the problem.
- 02
Validate feasibility
Our engineers analyse whether FOCUS™ AI can reliably identify the required deviations at production speed. Honest assessment — including the cases where we recommend a different approach.
- 03
Train FOCUS™ AI
Build and optimise the inspection model using representative production data — real defects, real variation, real production conditions.
ZERO DEFECTS IS THE TARGET.
ZERO RECALLS IS THE RESULT.
ZERORECALLS.AI
Vision that never loses focus.
Have an inspection challenge? Put our Vision AI to the test. We will tell you whether we can solve it — and if we can, we will show you how fast.
