Dimensional measurement with machine vision
Non-contact checks of lengths, diameters, hole positions and gaps on machined, stamped and moulded parts.
Technical review: Frank GuoReviewed
A camera station can measure lengths, diameters, hole positions, gaps and angles without touching the part, and flag parts outside tolerance. Its accuracy depends on field of view, camera resolution, optics, calibration, how the part is presented and the environment, so no single accuracy figure applies. We establish it on your own parts, against your tolerances and your reference measurements. It complements a CMM rather than replacing it.
What a vision measuring station checks
Four groups of dimensions that a camera measures from edges in the image.
Lengths and widths
Distances between edges, such as overall length, slot width or step position.
Diameters and radii
Outer and inner diameters, radii and roundness as seen in the image plane.
Hole positions
Hole and feature positions relative to datums on the part, and pitch between holes.
Gaps and angles
Gaps between assembled parts, edge distances and angles between edges.
What sets the accuracy
| Factor | Why it matters | How it is handled |
|---|---|---|
| Field of view and resolution | Each pixel covers a patch of the part; a larger view means coarser pixels. | Chosen from your tolerances; a large part may need several cameras. |
| Optics | With a standard lens, apparent size changes with distance from the lens. | A telecentric lens keeps size constant over its depth range, for parts at varying height. |
| Calibration | Converts pixels to units and corrects lens distortion. | Done with a calibration target and checked with master parts at agreed intervals. |
| Part presentation | Tilt or movement looks like a dimension change. | A fixture or datum stops hold each part the same way. |
| Lighting and edges | The measurement is only as good as the edge the software finds. | Backlight gives crisp silhouette edges; burrs and chamfers are agreed in advance. |
| Temperature and vibration | Parts and mountings expand and move. | A rigid mount and measuring parts at a known condition. |
Vision station and CMM compared
| Aspect | Vision station | CMM |
|---|---|---|
| Where it works | In the line, on every part. | In the metrology room, on sampled parts. |
| What it measures | Features visible in the camera view, mostly in one plane. | Full 3D geometry, including hidden and internal features. |
| Its role | Sorts every part against tolerance and shows trends. | Reference measurement and first-article inspection; checks the vision station. |
Where vision measurement stops
We agree these boundaries before a project starts. We do not quote an accuracy figure before testing your parts.
Not measurable with a standard camera
- Hidden features. Internal bores, undercuts and faces the camera cannot see.
- Height and flatness. A camera looking straight down sees one plane; depth needs a 3D sensor.
- Soft or unstable parts. Parts that deform or move during imaging do not give a stable result.
Works reliably when
- Parts sit the same way. A fixture or datum stops remove tilt and movement.
- Edges are well defined. Backlit silhouettes or clean, lit features.
- The tolerance is wide enough. The station's variation is small compared with the tolerance, as shown by the study below.
How a measuring station is laid out
Side view of a typical backlit station with a telecentric lens. The real layout follows your drawings and the sample test.
Swipe sideways to see the whole drawing →
What a pass and a reject look like
Illustrations, not measurements from a real part. Tolerances come from your drawing; how near-limit parts are handled is agreed with you.
Dimensions in tolerance
Both hole positions lie inside their tolerance zones relative to the datums.
Hole position out
One hole is shifted outside its tolerance zone. The part is rejected and the value logged.
Close to the limit
The diameter is within the station's own variation of the limit, so the part is diverted for a check on your gauge.
What we need to know about your line
These conditions decide the station design. Each is confirmed on site or during the sample test.
| Condition | What we need to know | Why it matters | Confirmed during |
|---|---|---|---|
| Drawings and tolerances | The critical dimensions, their tolerances and datums. | Decides field of view, resolution, optics and whether one camera is enough. | Design review |
| Part presentation | Conveyor, fixture, rotary table or robot; how repeatable it is. | Movement and tilt add directly to measurement error. | Site survey |
| Cycle time | Parts per minute or machine cycle time. | Sets exposure, trigger and whether parts must stop to be measured. | Site survey |
| Environment | Temperature swings, vibration, dust, coolant or oil on parts. | Affects calibration stability and the edges the camera sees. | Site survey |
| Reference measurement | Your CMM, gauges and measured master parts. | Needed to prove the station agrees with your accepted method. | Sample test |
| PLC and data | PLC make and model; whether measured values must be logged. | Decides reject handling and data output for trend monitoring. | Design review |
How accuracy is established and accepted
Measurement stations are accepted on evidence from your own parts, not on a catalogue figure.
Repeatability study
The same parts are measured repeatedly, with reloading, to show how much the station varies. The result is compared with your tolerance.
Agreement with reference
Parts measured on your CMM or gauges are measured by the station, and any systematic difference is found and corrected.
Calibration routine
A written procedure for checking the station with a master part or target, at an interval agreed with you, with records kept.
What to prepare for the sample test
We compare our results with your measurements before proposing a station. We confirm quantities and shipping with you first.
Sample preparation guideDrawings
Part drawings with critical dimensions, tolerances and datums marked.
Measured parts
Parts with your CMM or gauge results, including near-limit and out-of-tolerance parts.
Part variation
Parts from different batches, tools or cavities, as they arrive at the station.
Line information
Cycle time, how parts are handled, environment, PLC make and model.
Dimensional measurement questions
How accurate is camera-based dimensional measurement?
When are telecentric lenses needed?
Can a vision station replace our CMM?
Can it measure height or flatness?
Request an assessment for this application
Send your drawing and a few measured parts. We will tell you whether a vision station can measure them and how we would prove it.