Advanced License Plate Recognition with Squeezenet Effeciency

Abstract

License Plate Recognition (LPR) holds immense importance within the world of Intelligent Transportation Systems (ITS) due to its diverse applications. This study explores the implementation of a SqueezeNet model for LPR, addressing the critical role of intelligent transportation systems. By utilizing deep learning (DL) and image processing techniques, the purpose of the study is to enhance the accuracy and efficiency of automatic license plate recognition (ALPR). The characters on license plates are identified and recognized from a variety of vehicle images using the SqueezeNet architecture, which is well-known for its lightweight construction and computational effectiveness. The main steps in this study includes image processing to optimize the input quality, followed by character segmentation and recognition using the trained SqueezeNet model. The experimental results illustrates that the model achieves 99.65% of accuracy, thereby highlighting the model’s efficiency in real world applications such as automated toll collection and traffic monitoring. This study underscores the transformative potential of modern DL approaches in advancing ALPR systems.