“FALCON” Inspection Case 2: Crack Inspection of Drawing Press Processed Products
1．What kind of process did you use to consider visual inspection?
This system was designed to inspect the tiny cracks that occur during processing of draw-pressed metal parts, which previously had to be visually inspected on the order of several thousand pieces per day.
2．What was the challenge?
The part in question is difficult to inspect because it has a complex shape, defective parts are extremely small, and there are patterns caused by the material.
3．What Kind of System Did You Consider?
We have built an inspection system that allows inspectors to manually set the workpiece and turn on a switch to perform inspection and judgment.
4．How Did the Issue Go?
We created an imaging environment that makes it easy to see abnormalities, and used AI technology to distinguish between good and defective products in images that are extremely difficult to distinguish.
5．Why Did You Work With AFC Lab.?
Customers evaluated 5 to 6 visual inspection systems, including rule-based and AI, with the result that rule-based judgment was challenging. Therefore, multiple companies proposing AI visual inspection systems conducted several accuracy verifications. In addition, considerations were made from various perspectives, such as integration into automated devices. “FALCON” satisfied the inspection accuracy and uniquely met customer requirements only with our support in setting conditions, understanding manufacturing site constraints by our engineers.