By simplifying character reading applications, edge learning technology improves quality control and traceability.
Cognex Corporation today announced that its In-Sight 2800 vision system now features optical character recognition (OCR). Powered by edge learning technology, the new ViDi El Read tool seamlessly deciphers human-readable characters on reflective, non-flat, low-contrast surfaces.
"With edge learning technology, the In-Sight 2800 system redefines ease of use, allowing any user to deploy AI-based applications regardless of experience," said Lavanya Manohar, Cognex vice president of vision systems. For its part, the ViDi El Read tool replaces complex programming tasks with example-based training, which simplifies the configuration of optical recognition models to read characters on difficult surfaces, and even several lines of text simultaneously."
Thanks to its ease of use, OCR software, for example, allows food and beverage producers to easily read expiry dates, even on curved surfaces, in order to check the freshness of food and avoid product recalls from consumers. Pharmaceutical companies and medical device manufacturers can verify dates and lot codes to ensure vaccines are effective and meet current medical standards.
This solution also allows EV manufacturers to quickly locate and read laser-engraved alphanumeric text on the underside of their vehicle battery to improve traceability. Finally, logistics centres can decipher codes and texts on different types of packaging to ensure that products are transported correctly and avoid batch rework.
Implementing an OCR application is a tedious operation that can require several hours of programming by highly skilled engineers, preventing many companies from automating this type of inspection. With the ViDi El Read tool, models are directly configured and deployed to the device in minutes with just 10 sample images. The accuracy of the results is measured by a confidence score that is displayed under each of the characters formed, providing real-time visual feedback and ensuring reliable traceability throughout the supply chain.
Launched in April 2022, the In-Sight 2800 vision system was designed to solve complex classification applications. Other tools are under development and should be available in the coming months.