License Plate Recognition Using a Hybrid Deep Learning Based Model •
Year: 2026
Type: conference-paper
Abstract: Automatic License Plate Recognition System (AL-PRS) serves as a crucial component within Intelligent Transportation Systems (ITS). Accurate license plate recognition (LPR) is vital for modern intelligent infrastructure to effectively detect and identify vehicles. However, LPR remains a challenging task due to variations in license plate formats, inconsistent illumination, diverse viewing angles, occlusions, and deviations from standard templates. The objective of this study is to recognize license plates from vehicle images captured by CCTV cameras deployed across India. The proposed approach employs a Class Attention-Inception Network to detect vehicles in input images. Subsequently, morphological operations are applied to localize the license plate region within the detected vehicle, and a hybrid inception network is used for character recognition. Experimental results demonstrate a vehicle detection accuracy of 99.34% and a character recognition accuracy of 98.76%. The proposed system enhances surveillance efficiency, enabling faster and automated traffic law enforcement. Furthermore, it supports real-time vehicle tracking and monitoring, contributing to the development of safer and smarter cities. The model's scalability and adaptability also make it well-suited for deployment across diverse Indian traffic and environmental conditions.
Language: English
Subfield: Computer Vision and Pattern Recognition
Field: Computer Science
Domain: Physical Sciences
Open Access status: hybrid
OpenAlex ID: W7167222425