Enhanced Visibility of Vehicle License Plate Recognition Systems Using Adaptive Gabor Filters

Abstract

Automatic Number Plate Recognition (ANPR) is a critical technology within Intelligent Transportation Systems (ITS), particularly for traffic monitoring, law enforcement, and smart city development applications. Nevertheless, ANPR systems face significant challenges due to real-world conditions such as variable lighting, unclear license plate characters, and low image quality. These challenges are relevant in urban and highway environments where vehicles travel at 40 – 60 km/h speeds. This study introduces a novel integration of YOLO v8 for more precise license plate detection and adaptive Gabor filters for image enhancement, improving recognition accuracy under challenging real-world conditions using EasyOCR. The adaptive Gabor filters dynamically adjust their parameters to handle diverse image quality challenges, including uneven lighting, extreme viewing angles, and low resolution. Experimental results show that applying adaptive Gabor filters reduces the Character Error Rate (CER) from 7.5% to 4.98%. The result shows that the system enhanced ANPR system accuracy and reliability.

Access to Document

Other files and links

• Link to publication in Scopus

Fingerprint