Research On The Key Technologies Of License Plate Recognition

Along with the developing of traffic modernization, intelligent transportation system (ITS) won the best opportunity for development. The license plate recognition technologies (LPR) are getting intensive study as an important means of collecting information in ITS.In this paper, we investigate a variety of license plate recognition technologies at home and abroad. Combining with the license plate characteristics in the law of {The People’s Republic Of China Motor Vehicle Plate(GA36-2007)》. This paper proposed a new binary method, image segmentation method and image recognition method for LPR. Experimental results show that proposed methods can improve the rate of recognition than others.In this paper, we divide the license plate recognition system into three aspects: license plate localization, character segment and character recognition. This paper researched the key technologies of each part form the following aspects:(1) Research on the algorithms of license plate localization and calibration method. Standing out the license plate area through the image pretreatment and regional intensity, then select the license plate based on the calibration methods.(2) Research on the algorithms of character segment and calibration method. Disposing the license plate images by pretreatment and tilt correction, then though the methods of projection and calibration to segment characters.(3) Research on the algorithms of character recognition. Studying a variety of character recognition methods, including character recognition based on neural network, character recognition based on character structure and matching recognition based on the distance of Hausdoff.