Optical character recognition (OCR) technology started with the purpose to aid the blind – a computer to read text aloud. Today this technology is an essential part of every business for text recognition. An OCR application extracts and repurposes data from scanned documents, camera photos, and image-only PDFs. The software isolates letters on the image, turns them into words, and then turns the words into sentences, allowing access and alteration of the original material. It also eliminates the necessity for manual data entry.
However, the current OCR software in the market (i.e, Nanonets, OCRspace, Adobe Document Cloud, Amazon Textract, etc.) is still confined to scanning data from printed or handwritten documents in a specific uniform typeface. We at Fabrik are taking OCR to the next level, beyond papers and into the vision systems used for manufacturing across industries.
Machine vision systems are a collection of interconnected parts designed to use data extracted from digital images of the production lines and automatically direct manufacturing and production procedures like go/no testing and quality control processes. Errors or deviations in the production process are immediately identified and reported, enabling controlled adjustments to be performed in real-time to decrease scrap and eliminate costly downtime.
OCR systems built for embossed or real-world scenarios like factory floors heavily depend on complex hardware to perform recognition and verification (which makes it more expensive and harder to replace).
The Fabrik OCR system’s real-world effectiveness comes from large amounts of data used to train the recognition and flexibility of the system to move towards its needs.
Take for example, a tyre manufacturing industry - the embossed text on the tire needs to be verified before the tyre can be sent to another section. This verification needs to happen in multiple steps. If the OCR system was dependent on heavy hardware, it would cost a lot to set up inside the plant. But since Fabrik’s OCR system does not require heavy hardware setup, it can be set up and moved around with convenience.
The OCR system does not stop at just the verification. Feedback-based action can be set up through the system. E.g. If there is a specific error, the system can integrate into the communication devices to throw alerts, or if directly engaged in the IIOT systems, the system can automatically take actions based on that. The system augments the humans working on the factory floor.
Fabrik enables access to industry-leading computer vision algorithms to detect codes and sequences through our platforms. You can swiftly scan any code without any errors. The real-time dashboard captures every piece of information and displays the object in 3D for better visualization and error-proofing. The ability to transform the current management of error codes/symbols into an automated system thereby creating the first steps of quality digital twins is a powerful application for the automated management of key assets and processes for the future.