Can a person see the defect in a photo?
If a clear photo taken in good light does not show it, a camera will struggle too. If it only shows at one angle, that angle is the lighting design.
Get started
×Tell us about the line. A person reads it and replies.
Or talk it through first:
30-minute Google Meet call, Monday to Friday, 9 am to 5 pm Pacific.
Need a full study of the line first?
Guides
Most failed vision projects were decided before the purchase order: the defect could not be seen, the part could not be held still, or nobody agreed what a bad part was.
You do not need a camera to answer these. You need samples, a few photos of the line, and the people who decide what a reject is.
If a clear photo taken in good light does not show it, a camera will struggle too. If it only shows at one angle, that angle is the lighting design.
Consistent defects suit rule-based checks. Defects that change shape, size and position may need a learned model and many more sample images.
Line speed and spacing set the exposure time, the trigger and the reject timing. Gaps between packs matter as much as speed.
The camera, the light and a shield need space, often more than people expect. Note where a reject device can go and how far it is from the camera.
Quality, production and maintenance should agree on good, bad and borderline samples before anything is built. That set of samples becomes the test.
Bring 20 to 50 good samples and as many real bad ones as you can find, including borderline cases. They are worth more than any spec sheet.
FAQ
A short test on your real samples, usually on the bench, that shows whether the defect can be imaged reliably and what lighting and camera it needs.
As many real ones as you have. A handful is enough to start a rule-based check. Learned models need more, often hundreds.
Often the first test can, with shipped samples and line photos. A site visit follows before the design is final.
Then you find out before buying a station. Sometimes the answer is a different sensor, a change to the part handling, or a manual check.