Intelligent System for Vehicles License Plate Recognition Using a Hybrid Model of GAN, CNN and ELM

Abstract

The scientific community has given license plate recognition systems a lot of consideration. The current methods for vehicle identification need to be improved due to the swift increase in vehicle numbers. In order to lessen reliance on labor, a fully automated system is needed. With the growth of Intelligent Transportation Systems, demand for license plate recognition has increased significantly. License Plate Recognition (LPR) is susceptible to environmental factors such as a complex image background, angle view, and shift in illumination, it is still difficult to correctly recognize the digit letters on license plates. When reading license plates automatically, license plate recognition uses character recognition and image processing to identify the vehicles. The license plate detection and identification subsystems are typically combined into the vehicle license recognition system in order to locate the vehicle and identify the license plate. The Extreme Learning Machine (ELM) is used for categorization, identification, and training. This research suggests a GANCNN-ELM-based technique for detecting vehicle license plates. This method produces an accuracy of about 98.94% which outperforms the GAN-ELM, GAN-SVM, and GAN-CNN models.