Because we manufacture parts for different customers that look similar or even identical, our Planning and Quality teams identified part inspection as a process in need of an overhaul. Dedicated parts inspectors used hand-held equipment to visually evaluate each part, a process with a high potential for error.
To mitigate the risk of misidentification, we partnered with Akeratos LLC to create an automated solution that drastically reduced the potential for human error in the parts identification process. We developed a Vision system that uses a camera and computer software to inspect and identify each part. Parts are programmed in the Vision system to identify key features that make each part unique. Programming the parts in our Vision system begins by capturing a clear picture of each part, which takes into consideration the best possible orientation and lighting conditions.
Once the picture is taken, the software is programmed to recognize unique features that allow the differentiation of part configurations. These unique features can include area widths, hole locations, hole sizes, or other features. After the program is written, parts are tested using the software to determine if the unique features on the part match what the software knows is a good part. The operator receives a “GOOD” or “BAD” indication from the machine, which quickly tells the operator if the part is the correct configuration.