Lighting and presentation are part of the vision system.
A high-resolution camera cannot compensate for uncontrolled reflections, variable shadows or a product that cannot be distinguished from its background. Trials use representative products and expected variation to establish contrast, field of view, exposure and processing margin.
Robot guidance also needs an accurate relationship between camera, product and robot coordinates. Calibration, conveyor encoder data and mechanical stability are designed and verified as part of the cell—not treated as a software afterthought.
Locate and orient
Calculate position and angle for robot picking from trays, fixtures or moving conveyors.
Inspect and verify
Check presence, shape, assembly, surface features or other measurable criteria before or after handling.
Identify
Read supported codes, marks or features to select a recipe or route the product.
Reject and recover
Define what happens to uncertain, failed or unpickable products without stopping the line unnecessarily.
Vision system design questions
The camera is one element in a controlled image-acquisition and decision process.
| Object and task | Feature to locate, inspect or identify; acceptable variation; pass/fail or measurement rules |
|---|---|
| Optics | Field of view, working distance, resolution, depth of field, lens and enclosure |
| Lighting | Geometry, wavelength, diffusion, polarisation, strobing and control of ambient light |
| Triggering | Sensor or encoder trigger, exposure timing, object motion and image availability |
| Calibration | Camera-to-robot relationship, conveyor tracking, datum stability and re-calibration method |
| Decision handling | Confidence thresholds, rejects, retries, audit data and operator diagnostics |