Machine vision based fill control system for blister packing machine


This is a machine vision based inspection system, designed to detect and reject faulty blisters in a blister packing machine. We have installed this system on a Bosch blister packing machine which originally had an out of date photodiode array fill control system. Our fill control system uses sophisticated image processing techniques to detect missing and not desired types of  capsules in real-time.




The system consists of a color CCD camera, a framegrabber card, PC and a digital IO card. The framegrabber is PCI-1405 of National Instruments and DIO card is designed and made by KDI. We used LabVIEW and Vision, both from National Instruments to develop a robust, real-time image processing software which can accurately process medium and even low quality images to detect several types of blisters and capsules.




As the processor prices are much lower than high quality optics and cameras, we used the processing speed of an Intel Pentium Dual Core processor to execute our extensive amount of codes in a relatively short time to manage to process each pair of blisters, synchronized with the packing machine's pace. With doing so, our system works excellent with low cost CCD camera and lighting boards.

The software also controls two electrical valves to shoot faulty blisters in the waste baskets using high air flow.


System features:

The software is optimized to use the maximum power of dual core processors

Full support of the Persian language  

Generating HTML reports of the system's operation, including  number of inspected, correct and faulty blisters

Integrated database