A full analysis of the packaging of products moving on the conveyor belt
We provide a vision system for industrial automation that can be installed with classic belt conveyors. Using system of 5 cameras, the algorithm recognizes and reads desired content – such as serial numbers, expiry dates and bar codes. Thanks to the use of specialized lighting placed in 3 separate chambers, it is possible to accurately recognize both traditional laser and needle-printed inscriptions, as well as embossed inscriptions, often difficult to read with the naked eye. Text identification is carried out using a group of modern computer vision algorithms and deep learning, performed on an integrated computing unit. It is also possible to send information read from the boxes to the client's IT systems.
Entering serial number, expiration date and bar code into the ERP system takes about 30 seconds for a human worker. Our system only needs 5 seconds, and it works around the clock, processing over 17,000 products a day.
The dedicated lighting system and 5 industrial cameras scan each side of the product while it passes through the device, taking photographs in high resolution.
Reading algorithms are prepared for every type of font, including economical needle printing and embossed inscriptions that require appropriate lighting to be visible to the naked eye.
The solution uses the latest achievements of machine learning and computer vision, currently revolutionizing the field of optical character recognition (OCR).
[...] AI Factory provided a solution based on artificial intelligence that successfully addressed a real problem. All requirements were met and completed on time. [...]