Height and depth
Measure steps, recesses, protrusions and component seating against agreed limits.
Height, depth and profile measurement
3D machine vision inspection systems
Measure shape as well as appearance. We integrate 3D sensors, product handling, inspection software and the line response around the feature that must be controlled.
Application-led design Proved against the real product, defect and production conditions.Direct answer
When does an inspection need 3D machine vision?
Use 3D inspection when the pass/fail decision depends on height, depth, profile, flatness, volume or surface geometry that a conventional 2D image cannot prove reliably. The sensing method is selected after the required measurement range, smallest defect, material, speed and product presentation are understood.
Typical project scope
What the system can cover
Each requirement is tested against representative products, normal variation, line speed and the agreed production response.
Gap and flush
Inspect visible gaps, offsets and relative surface positions in assemblies.
Profile and contour
Compare cross-sections, edges, beads and formed profiles with an approved tolerance window.
Flatness and warpage
Identify out-of-plane variation where the field of view and sensor resolution support the requirement.
Area and volume
Calculate geometry-derived area or volume for presented products and deposited materials.
3D surface defects
Detect agreed dents, raised material, chips or deformation that may be difficult to separate by colour alone.
Engineering factors
What determines a dependable result?
Choose the sensing method
Laser profiling, structured light, stereo and other 3D methods have different accuracy, speed, working-distance and surface limitations.
Prove the material
Gloss, transparency, very dark surfaces, vibration and steep edges can affect the returned height data and must be tested on real samples.
Connect the measurement
Calibration, product tracking, recipe control, reject timing and retained results determine whether the measurement becomes a dependable production control.
Production outcome
Turn the inspection into line control
The useful result is a repeatable decision that operators can understand and the production line can act on.
- Measure every presented component rather than relying only on periodic samples
- Separate shape faults from harmless colour or print variation
- Detect assembly drift before a whole batch moves downstream
- Link an out-of-tolerance result to rejection, alarm, stop or process feedback
Practical delivery
From requirement to accepted system
Define the geometry
State the feature, tolerance, field of view and smallest unacceptable variation.
Capture representative parts
Test good, failed and borderline samples across normal material and finish variation.
Prove repeatability
Challenge speed, vibration, presentation, calibration and environmental influences.
Integrate and accept
Connect recipes, tracking and response, then test against agreed production evidence.
Decision guidance
2D and 3D inspection solve different problems
A 2D camera is often the simplest choice for position, colour, print and visible edge checks. A 3D sensor adds measured height or profile information when contrast alone is ambiguous. Some applications benefit from both: 2D verifies appearance and identity while 3D proves geometry.
- Use 2D where stable contrast provides the required evidence.
- Use 3D where the acceptance criterion is genuinely geometric.
- Specify the smallest relevant feature, not only the overall product size.
- Agree calibration checks and the response to loss of valid measurement data.
Related guidance
Continue planning your inspection
Questions, answered
3D machine vision inspection systems FAQs
What is the difference between 2D and 3D machine vision?
2D vision evaluates image position, shape, colour and contrast in a flat image. 3D vision adds height or depth data, allowing geometry such as steps, profiles, volume and warpage to be measured.
Can 3D vision inspect reflective or dark parts?
Often, but performance depends on the sensor method, angle, finish and required resolution. Representative samples must be tested before the architecture is fixed.
Can 3D inspection run in line?
Yes, when acquisition and processing fit the line cycle and product movement is controlled. Speed, vibration and trigger repeatability are part of the feasibility test.
Can a 3D system be retrofitted?
Frequently, subject to working distance, access, guarding, product presentation and available reject or stop distance. A line survey confirms the practical constraints.
How accurate is 3D machine vision?
Accuracy is application-specific and depends on sensor resolution, measurement range, calibration, surface behaviour, mounting stability and environment. It should be stated and proved against the real feature and tolerance.