Research And Realization On Character Segmentation Methods For Vehicle License Plate

License Plate recognition system is one important topic of the Intelligence Traffic Management. License plate recognition (LPR) has great application prospective in tolling, monitoring traffic flow, investigating on-the-spot traffic accident, handling breaks of the traffic regulations and keeping parking secure. People pay increasing attention to LPR based on digital image processing and pattern recognition.License Plate recognition system is an important subject to be studied in the intelligent traffic. It is mainly composed of three keyparts:license plate location license plate character segmentation and license plate character recognition. License plate character segmentation,as a key part of the license plate recognition system, is mainly studied in this paper.There are three key points in the license plate segmentation system. They are binarization, tilt correction and character segmentation.In the binary algorithm, the thesis has combined with advantages of the two global binarization Otsu algorithm and partial binarization Bernsen in order to adopt a new method of threshold segmentation with license plate features—Four Section Block Binary Algorithm.In the tilt correction algorithm, the thesis has adopted the tilt correction algorithm, combining with linear fitting and Radon Change for horizontal tilt correction. However, for vertical tilt correction, the thesis has applied character projection to get vertical tilt angle and calibrated the title by the method of coordinate transformation. It can avoid losing much character details, be helpful to consequence segmentation and can improve identification rate.In the character segmentation algorithm, the thesis adopted a method of license plate character segmentation with connectivity, vertical projection and template matching. Firstly, the character's width has been found by the connectivity; Secondly, separator location has been found, according to the projection; At last, the segmentation has been carried by the template which has been set up by connectivity character's width. Then, single character has been adjusted microscopically by projection. The author proposes the corresponding segmentation method for each type of license plate, after considering fully existing license plate types, especially for double license plate, police and consular license plates etc.License plate segmentation system is self-adaptive. It could process the images that sampled from different environment, and get a wonderful result.