Automatic Bengali number plate reader

In this paper, we propose connected component analysis based algorithm to automatically detect and read Bengali number plates used in the metropolitan cities of Bangladesh. Our proposed automatic number plate recognition (ANPR) system involves image pre-processing and morphological operation followed by edge detection, regional localization and character segmentation to identify the Bengali characters in the number plate efficiently as well as with less computational complexity. For various weather conditions, our proposed algorithm shows a detection accuracy rate of ∼95% with an average processing time of 0.75 seconds. Our system might be highly effective in real time traffic control, security enhancement as well in electronic toll collection.