The application of automatic systems is being promoted in Myanmar. At the same time, the number of vehicles including private cars, taxies and buses are increasing day by day, automatic vehicle security systems are necessarily important to increase national security in both public and private places. License Plate Recognizing (LPR) plays an essential role in automatic vehicle access control systems which are popularly used for private and public security today. LPR is based on image processing technology. Therefore, to develop a reliable LPR system for recognizing Myanmar vehicle license plate is currently in demand.
In this work, an image processing method for localizing and recognizing Myanmar vehicle license plate is developed. It consists of three main steps; license plate localization, character segmentation and character recognition. The localization of license plate is performed based on Euler number and license plate aspect ratio. Then, boundary tracing method is used for character segmentation. Finally, the character recognition is implemented based on partially cut character structure.
The performance of proposed license plate recognition method is tested for real field Myanmar vehicle images including front view, back view, skewed view, near view and far view. The proposed method shows a reliable performance by detecting even in skewed plates without skewness correction. Thus, this work contributes a LPR system which is practically useful in automatic vehicle security system in Myanmar.