Challenge: Recognize the type of battery and examine its cells for orientation, integrate solution with an existing PLC controller.
Difficulties: 40 different battery types are manufactured, integrate solution in a hostile dusty and corrosive industrial environment with unstable power supply and high levels of electromagnetic (EM) interference.
Solution: The product type can be determined from its dimensions. Edge detection reveals objects contour then its dimensions are measured. Inserted battery cells are collored red on one side. Region of interest is therefore divided in several areas. Each of them is tested for presence of red colour giving us the information about battery cell orientation. National Instrument Compact Vision System offered a robust dedicated RT processing unit that would withstand all of the ambiental issues. Special shilded firewire cables with ferrite beads attenuate EM interference, UPS power supply protect the equipment from electrical spikes, power supply fluctuations and faliures etc.
Result: Our solution discoveres all human errors caused during manual insertion of battery cells before the product is hermetically sealed and is rejected by a final electrical test. This early enough stage of error detection helps save production resources considerably.