Research, Image Processing And Neural Network-based License Plate Recognition System

License Plate Recognition System (LPR) is the core component of Intelligent Transportation Systems (ITS). LPR plays an important role in a modern traffic management system. In recent years, the research on the license plate recognition system has become a hot issue.First of all, the paper introduces the research status on license Plate recognition system, and carry out a comprehensive exposition on four key modules about image Pre-processing, license plate location, character segmentation and identification.Give an in-depth study of character recognition technology, design and achieve a license plate recognition system. In order to improve the recognition accuracy and robustness, give a careful analysis of focus and difficult problems which are encountered in the course of four modular designs, and propose some solution.In the design of pre-processing module, use a variety of image processing technology and adopt a series of pre-processing algorithm including fuzzy c-means binarization, Morphology filtering,Sobel edge detection operator.The paper study and compare the methods of the feature-based statistics, the improved sobel edge detection, particle image correlation method, and experiments prove: particle image correlation method is the most good optimization program.In the license plate character segmentation module, present a template matching method which based on multiscale segmentation algorithm, the use of scaling methods to find the global optimum plate region template matching information, combining the Hough transform this algorithm has good robustness. In the light of the current study, analyze the status and quote BP networks, give a set of additional momentum method to improve the BP algorithm. The comparing experiments verify the effectiveness of the improved algorithm.Experiments show that the methods in this paper achieve better recognition results, with a certain degree of efficiency, robustness and real-time.