Optical character recognition is widely used to recognize characters from a variety of sources such as printed pages of text, license plates, serial numbers, and product codes. It is basically the conversion of text/characters into a digital format that can then be used to electronically transmit that information, make it searchable or resulting in processing such as sorting items or sending out tickets.
Post offices in the United States and the United Kingdom were just a couple of the early users that adopted machine vision systems to convert postal addresses on envelopes to digital text. Of course there is then a mechanical process triggering sorting of mail and rejecting addresses that cannot be completely read by the machines.
OCR can also used to read license plate numbers from images of cars caught by red light cameras or speed trap cameras. Some cities in the United States are using license plate readers for traffic control and parking monitoring to track vehicles and monitor how long they have been parked in a parking spot.
Text to speech is another popular application for OCR although the results do not sound like a natural reader it is still useful in situations where conversion is required.
Recognizing handwriting is a much more complex and difficult task to accomplish using technology – I can’t read my own handwriting sometimes! But researchers out there are trying to figure out a way to do that too. Just might take a while.
- Parm Dhillon
 |
Sample of Sudeep's handwriting that I decode frequently.
Handwriting is still a challenge for OCR applications. |
Labels: license plate reader, Optical character recognition, product code reader
