Advanced Image Dehazing Techniques for Enhanced Text Extraction from Vehicle Number Plates for Traffic Monitoring System

In this paper, we have proposed guided filter-based image dehazing techniques to convert the hazy Indian vehicle license plate and generate the dehazed image of the same. The output of the proposed dehazing technique is inserted into the Optical Character Recognizer (OCR) to extract the text as a final result of the proposed outcome. To establish the efficiency and effectiveness of the proposed approach we have implemented our algorithm not only with the hazy Indian vehicle license plate (IVLP) dataset but also with the help of the non-hazy Indian vehicle license plate dataset. Our proposed method is compared with existing recent IVLP detection algorithms, both quantitively and qualitatively using various performance parameters such as MSE, RMSE, SNR, PSNR, Accuracy, and SSIM. Comparative analysis shows that our proposed method overpowers the existing methods described in the survey table. Our results demonstrate significant improvements in text extraction accuracy, showcasing the potential of advanced dehazing techniques in real-world applications. This paper also gives a clear idea about the importance of clear visibility of vehicle number plate recognition in adverse weather conditions such as haze, fog, and rains as well as to avoid accidents on roads.