License Plate Character Recognition Algorithm Based On BP Neural Network

License Plate Recognition (LPR) technology is one of the most important parts in Intelligent Transport System (ITS), including License Plate Location, Characters Segmentation and Characters Recognition.The neural network pattern recognition is one of the important research areas in the field of pattern recognition recently. The non-linear and ability of self-study and self-organization make it have unique predominance.The paper chooses the recognition method based on Back Propagation (BP) neural network, and selects several different features of characters as the neural network input. Trains the neural network, achieves the license plate character recognition and compares their advantages and disadvantages. Including KL transform based on gray image features, KL transform based on binary image features, and VTDHTD features. Finally, respectively counts the accuracy and time of several identification methods, and proposes improved methods:correct the results of recognition.In the Intelligent Transport System, because people require a more accurate license plate recognition system, it will have some significance on theory and practice that combine the neural network with the LPR.