License Plate Recognition System Using Computer Vision

The license plate recognition system plays a key role in traffic system of any country. It can also be used as a security measure in highly sensitive zones. Many LPRS have been developed by various developers in order to detect or read the license plates of cars. This LPRS (License Plate Recognition system) combines the power of OpenCV, Imutils and Pytesseract as it offers cost-effective and efficient solution. Here we are using pytesseract to convert optical image into a string while OpenCV and Imutils is used for the localization of license plate. Firstly, the original picture is converted into grayed picture which is then converted into smoothened image. To detect number plate in a car it converts image further into edged one to find contours. Our system had achieved 87.5% accuracy in localizing the targeted area which is number plate of car and 62.5% accuracy in character recognition.