License Plate Detection and Recognition System based on Morphological Approach and Feed-Forward Neural Network

License Plate Recognition (LPR) is a mass superintending system that captures the image of vehicles and recognizes their license number. There has been a little of work done on Bangla license plate detection which is very important for recognizing the Bangla license plate. The plates with different backgrounds make it more complicated to use the existing algorithms. In this paper, we introduce a new algorithm for detecting Bangla license plate. At first, vehicle location is determined. In Bangladesh, Bangla license plates have some of the unique colors like green for the commercial vehicle and white for the personal vehicle. That’s why the portions of green and white colors are selected with the matching RGB intensity of the plate. Then contour algorithm and aspect ratio have been used to locate the license plate region. The rows of the license plate contain registration information that are separated using horizontal projection with the appropriate threshold value. After that, the characters and the digits are also separated using vertical projection with the same threshold value. Finally, we recognize the characters and digits with the help of back-propagation feed-forward neural networks. The proposed algorithms have been tested by 180 images taken under different conditions. The success rate of the license plate detection, segmentation, and recognition process is 93.89%, 98.22%, and 92.77%, respectively.