Research On License Plate Recognition System Based On Wavelet Transform

Recent years, more and more people have thrown themselves into the research ontheory and application of ITS (Intelligent Transportation System) in the world. Althoughthere are lots of products in the intelligent transportation field, there remains manyproblems to be solved, because of human factors and ceaseless change of the objectiveenvironment. Thinking of this point, the article makes a special effort to research licenseplate recognition (LPR), which is core content of ITS. Using wavelet analysis to LPRsystem is proposed to optimize performance of algorithms of license plate location andcharacters segmentation. And combine with BP neural network to identify the licenseplate characters.Firstly, this article studies the relation of wavelet analysis and image processing andthe application of wavelet transform in the LPR system. And put forward a algorithm oflicense plate location based on wavelet transform of G component. Combined withmorphological operation, the algorithm could position license plates accurately, whichare in different environment.And it’s good practicability and robustness. Secondly thispaper emulates the algorithm of image denoising based on wavelet transform, and thisalgorithm can remove the background noise and reduce character adhesion. At last, thispaper studies the algorithm of character segmentation based on wavelet transform andvertical projection diagram, this algorithm can syncopate single character accuratelyfrom license plate image.Besides, this paper studies the algorithm of characters recognition based on BPneural network and achieves a high recognition rate, and it’s good practicability androbustness.