Dark Channel Prior (DCP) based Bangla car plate detection and recognition in foggy weather

Abstract

"The intention of autonomous license plate (LP) detection is to find an LP spot in a picture without any kind of human intervention. Automatic license plate detection (ALPD) techniques have been developed; however, the majority of them do not take into account the numerous potentially hazardous image cases that frequently arise during actual driving scenarios. Hazardous picture conditions include low contrast settings, objects that are comparable to LP in the background, LP areas that are inclined horizontally, and weather effects like rain or fog. In this study, we present a novel approach to localizing and recognizing Bangla car plates in foggy or rainy weather using the Dark Channel Prior (DCP) approach for dehazing foggy images. There are three parts to the suggested method. In the first stage, the DCP dehazing algorithm is used to reduce the foggy effect on the input images. In the second stage, a YOLOv8 object detection model is used to detect the Bangla license plates from the dehazed images and lastly, the OCR technique is used to recognize and extract texts from the identified images of the license plates. Thus, our study aims to use DCP, the YOLOv8 algorithm, and the OCR technique to identify and recognize Bangla vehicle plates in foggy conditions in order to improve transportation safety, law enforcement, traffic flow, and tax revenue collection."