Research On The Image Recognition Technique For Moving Vehicles

In order to promote the development of Intelligent Transportation System, this paper researches the methods of vehicle classification and number plate recognition based on image process.In the paper, vehicle classification technique is firstly discussed. Some technologies such as image frames difference, histogram processing and median filter are applied to realize vehicle classification recognition module, and features extracted from the vehicle lateral image are put into BP neural network, which can classify the vehicle. Vehicle number plate recognition module is also discussed in the paper. Number plate image process consists of following several steps: number plate location, image binarization, plate contrast analysis, plate characters segmentation, feature extraction and three kinds of neural networks are used to realize characters recognition.The methods proposed by the paper are used to process actual vehicle images and the experiments results show that these methods are right and effective.