Fault-tolerant Certification License Plate Characters Recognition Research And Implementation Of The Algorithm

License plate recognition has intensive application prospects, but currently some of the key technical level needs to be improved. This paper is based on the National Natural Science Foundation project on "artificial brain information processing new neural network model" No.60673101and the National Natural Science Foundation project on "artificial brain with desired fault tolerant domain associative memory new neural network model" No.60973048, learning from the fault-tolerant authentication recognition ideological which is proposed by my mentor, has researched the key technical problems about image pre-processing, plate location, character segmentation and character recognition. It proposed an effective way to location which is use the color and shape information of the license plate and license plate characters fault-tolerant authentication recognition algorithm based on BP neural network. The algorithm including the license plate standard character feature determination algorithm, the Hamming distance between standard characters calculations, the standard characters fault-tolerant determination and the improved BP neural network which is trained by all samples on the fault-tolerant domain.The license plate character fault-tolerant authentication recognition algorithm can theoretically ensure that the domestic license plate character in the feature space on the fault-tolerant domain radius is2, and in the fault-tolerant domain the correct recognition rate is100%.