Orientation checks
Confirming top/bottom, leading feature or rotational position.
Inspect, classify and reject
Vision inspection & sorting for reliable part presentation.
Camera and sensor-based sorting systems that check orientation, presence, geometry, colour, print or selected defects before parts enter the next process.
- Sample-led design
- Defined orientation
- Sustained output
- Integrated hand-off
How it works
What is a vision inspection & sorting?
Vision sorting combines controlled presentation, lighting, cameras, image processing and rejection logic. It is most reliable when the inspection task, product presentation and reject handling are engineered as one system.
System overview
Designed around the real production condition.
The inspection station can be integrated with bowl or flexible feeding, conveyors, trigger sensors, reject devices, confirmation sensors, reject bins, line controls and data reporting.
- Defined inspection criteria linked to physical samples
- Lighting and optics selected around material, finish and feature size
- Pass, reject, rework or classifying outputs available
- Image records and production data can be included where required
Where it fits
Applications that merit assessment.
Presence and assembly
Checking holes, inserts, clips, seals or component completeness.
Classification
Separating variants by geometry, colour, mark or code.
Engineering decisions
Performance is determined at the interfaces.
The following issues are reviewed before the final tooling and controls are fixed.
What is objectively visible?
The inspection must be translated from human judgement into measurable image features.
Presentation repeatability
Stable position and controlled lighting reduce false decisions.
Reject verification
A reject command should be paired with detection or controlled containment where risk requires it.
Project definition
Information we need for an accurate proposal.
Drawings help, but representative production parts and the required hand-off condition are usually decisive.
| Defect library | Good parts, known defect samples and the smallest acceptable/rejectable feature. |
|---|---|
| Presentation | Part orientation, motion, spacing and camera access. |
| Inspection rate | Sustained parts per minute and exposure constraints. |
| Decision outputs | Pass/reject, multiple grades, line stop or data record. |
| Traceability | Image retention, counts, batch data and system connectivity. |
| Validation | Challenge samples, repeatability test and agreed false accept/reject methodology. |
Acceptance planning
Agree the test criteria before manufacture.
Agree the component set, sustained accepted output, test duration, normal interventions, orientation criteria, surface quality and downstream interface. This converts a demonstration into an acceptance test.
Common questions
Vision inspection & sorting questions.
What defects can a vision sorting system detect?
It can assess features that are visible and sufficiently repeatable under controlled imaging—for example orientation, presence, dimensions, colour, print and certain surface defects.
Can a camera guarantee every cosmetic defect is found?
No inspection should be specified in vague terms. The detectable defect type, size, contrast, location and sample variation must be agreed and validated.
How are rejected parts handled?
Options include air rejection, diverters, pick-and-place, gates or stopping the line. The method depends on product behaviour, speed and containment risk.
Can images be saved?
Yes, subject to storage, network and traceability requirements. Systems may save rejects, statistical samples or every inspection.
Can AI vision be used?
Machine-learning tools can be useful for variable visual defects, but they still require representative training data, validation and a controlled production process.
Related systems