Type-2 Fuzzy Based Tesseract Text Relative Recognition Algorithm for Number Plate Identification

Abstract

Automatic number plate authentication is a mass tracking system that captures vehicles' images and authenticates their license number. Number plate detection can be used in many applications such as electronic payment methods (toll payment, parking payment), freeway, arterial monitoring systems for traffic monitoring, and stolen vehicles. Several methods have been proposed in previous studies to solve the following challenges (location size, size, color, font, opacity, gradient, etc.) by retrieving information from the captured image. However, each method has its drawback between the accuracy of the detection and the processing speed. This ambiguous inference system only gives an estimate of the limit plate of the clear hypothetical system answer. These methods work at different times for a different number of models. The Type2 Fuzzy based TesseractText Relative Recognition Algorithm (TTRR) obtains the image dataset from the JSON (JavaScript Object Notation) file and processes it through the link in the dataset. In the proposed detection, the method uses ambiguous logic classifiers that help identify plates and characters with excellent accuracy and processing speed. Compared to many significant contributions, Type2 Fuzzy based TTRR yields better results.