License Plate Recognition Based On Modularization Data Fusion

License plate recognition is an important part of the traffic charging on high way. We analyze the system license plate recognition in detail and test the location of license plate, the colour confirmation of license plate and the character recogntion of the license plate. And we investigate the recognition of the type of the vehicle in particular.After describe the principle and application of Data Fusion, introduces its hierarchical structure model, and briefly showed the implementation on every layer's, the paper analyse the present situation of traffic charging on high way and the difficulties of the license plate recognition in China. Depended on the feature of license plate of vehicle, this paper inosculate the multi-sensors Data Fusion scheme using back propagate algorithm and the color feature of license plate to recognize the type of vehicle.When use the multi-sensors Data Fusion scheme, the accuracy of reconigition is 90.5%. And based on the color of the license plate of vehicle, the accuracy of reconigion is 82.4%. While the last data fusion based on decision level layer, the accuracy is 91.9%. It is a higher accuracy compared with each single one.