The Research Of Vehicle Plate License Auto-Recognition

IN 1990's, The Vehicle Plate Recognition research was set out. With the promotion of cross-science field, the Vehicle Plate recognition technology has made rapid strides. The Vehicle Plate recognition technology belongs to the category of pattern recognition, and is also closely associated with digital image processing. The Vehicle Plate recognition has a significant meaningrulness for ITS(Intelligence Transportation System), ETC(electronic toll collection) , regional vehicle administration and depot vehicle administration. It becomes an important research subject of information-processing technology.At first this paper introduces a background and a theoretical base of the Vehicle Plate Recognition system, including pattern recognition, basic theory of image processing, ANN (artificial neural network) and fundamental knowledge of wavelet. With above basic knowledge and theories, research of the Vehicle Plate Recognition has been completed, which is made up of there main steps: plate extraction, plate character segmentation and character recognition. After the analysis of these steps, in this paper, we advanced plate segmentation based on wavelet texture segmentation, plate character segmentation based on part-projection segmentation method and character recognition based on multi-group ANN recognition. There is a scheme of Vehicle Plate Recognition system and model realizations of these steps.In the part of this system's design and realization, the detail of a design and realization project is available. When designing this system, we set apart several modules from the main system, and brought forward requisition for the system's functions, interface designs, and designs of some crucial arithmetic, and we give details in the project-developing environment "Matlab" .in the last part of this thesis, we give the merit of our algorithm.