Research on Car Emblem Recognition Algorithm
|School||Liaoning Normal University|
|Course||Applied Computer Technology|
|Keywords||Car emblem recognition Knowledge Systems Neural Networks Inference engine Manifold Learning|
With economic development , the number of vehicles increases sharply , traffic problems have become increasingly prominent , which makes intelligent transportation system has become a hot research field , by the increasingly widespread concern . Vehicle identification is one of the important research topics in the field of intelligent transportation applications , is the core component of the intelligent transportation system , and can be widely used in dealing with traffic accidents , illegal vehicle monitoring , vehicle management , such as parking , airports , ports , residential entrance very broad application prospects . The vehicle identification information of the need to maximize the use of the vehicle to confirm the vehicle on a road . But only used the license plate and vehicle information , vehicle there is a more important message is that the car emblem information . The car emblem are the important information of the vehicle , is an iconic image of the vehicle , it contains not only the vehicle information, it is more important is that also contains the information of the manufacturer and it is difficult to replace . The combination car emblem , license plate and vehicle information , and will greatly improve the reliability of the vehicle identification . The car emblem picture carefully research a car emblem based on knowledge of the neural network recognition algorithm and a study based on the popular car emblem recognition algorithm . The car emblem recognition algorithm based on knowledge of the neural network as the theoretical framework of the theory of knowledge systems , is composed of two parts of the knowledge base and the inference engine . Knowledge base is used to store the used car emblem recognition knowledge , this paper to extract the car emblem gap and contour features for BP neural network training , sample car emblem contour feature , after a the car emblem gap number of characteristics and training The neural network weights to build the knowledge base . The inference engine is the use of knowledge in the knowledge base of the car emblem identify . Experiments prove that this method is not only to identify fast and high recognition rate . The car emblem recognition algorithm based on manifold learning is the picture of the sample , and to be recognized car emblem data dimensionality reduction processing , the car emblem picture identification operation in low-dimensional space , thus greatly reducing the computational amount of data use pictures of the car emblem texture information . The experimental results demonstrate the feasibility of the method .