Technology Research On AGV Traffic Sign Recognition Based On Image

With the development of urbanization, the popularity of cars, the problem which contents the road congestion, the frequent traffic accidents, the low efficient transport is more and more serious. Under the background, many developed countries began to use high-tech for changing the existing traffic system - the intelligent transportation system that is integrated information management system integrating communication, dection, control and computer. The key technology for realizing the system is developping smart vehicle with initiative security technology.Computer vision-based traffic sign recognition is smart to drive one of the key technical and difficult. Thus, in the automotive vision systems, how to effectively identify road signs is a very important research topic. At present, there are a variety of traffic sign recognition algorithms with different features ,which play a certain effect in certain occasions, but there are also some drawbacks. The research project aim is to overcome these deficiencies and improves the real-time traffic sign recognition algorithms and accuracy, which plays a great significance in practical applications.In the analysis and summary of the traffic signs at home and abroad to detect and classify on the basis of recognition, the text studies four key technologies e traffic sign recognition: pre-processing techniques, segmentation, feature extraction and neural network classification techniques, and proposes traffic signs recognition system framework. Then conbines the specific target identification and the segmentation and feature extraction techniques have been done a thorough analysis, proposing based on moment invariants and fuzzy wavelet neural network traffic sign recognition algorithm.Put the fuzzy technology into neural network,and using the advantages of fuzzy techniques to overcome the shortcomings of neural networks , finally discussed in detail based on moment invariants and fuzzy wavelet neural network identification method and BP neural network with experimental comparison, thus proving based on moment invariants and fuzzy wavelet neural network traffic sign recognition algorithm is moe effective and more superior. Experiments show that the proposed algorithm in this paper compared with the traditional neural network algorithm ,its training speed is faster, and the accuracy is higher ,and it can meet requirements about intelligent driving system in the road traffic sign recognition.