BanglaTag: Next-Gen Bangladeshi Number Plate Recognition Hierarchical Processing Network

Bangladesh is a nation experiencing rapid economic growth, resulting in a notable increase in the number of automobiles and the need for sophisticated infrastructural solutions. ANPR systems are essential for optimizing transportation systems in a variety of scenarios. Previous research on Bangladeshi ANPR systems has faced challenges with low accuracy and efficiency, mainly because of the presence of Bengali characters on license plates. In response, we propose an advanced end-to-end sequential deep learning framework. Our method consists of three main steps: initially isolating the number plate area from visual media, then separating the characters on the plate, and then utilizing optical character recognition methods for character identification. Our methodology significantly enhances accuracy and efficiency by utilizing numerous neural networks in a sequential approach and fine-tuning the models, surpassing previous methods. Our technology significantly decreases the need for computational resources and time, all while delivering exceptional performance. The research makes a distinctive contribution by providing a thorough answer to the issues presented by Bengali characters in ANPR systems, delivering a precise and effective solution tailored for the Bangladeshi environment.