License Plate Detection using YOLO v8 and Performance Evaluation of EasyOCR, PaddleOCR and Tesseract

License plate recognition is critical to the functioning of the control and surveillance systems. Vehicle identification is based on number plates, which are made up of a certain arrangement of letters and digits. However, manually identifying every number plate is a laborious and difficult task. The majority of earlier automatic number plate recognition (ANPR) systems were restricted in how they could operate. These conditions included being limited to fixed illumination, background noise in the vehicle image. The primary goal of this project is to develop a reliable number plate recognition model that functions with the help of YOLO v8 for license plate segmentation followed with a comparison model of EasyOCR, PaddleOCR, Tesseract models based on best model for character recognition of license plate. The steps involved include processing of vehicle image, segmentation of license plate from the image, followed by internal segmentation of characters on the plate and pattern matching. Data pre-processing is finally performed in order to read and process the license plate. Problems like background noise in image like characters apart from the license plate and number-based characters being wrongly predicted as alphabets have been tackled.