The Key Research Of Vehicle License Plate Recognition Under Natural Environment

Vehicle license plate automatic recognition is an important application of computer vision and pattern recognition in the intelligent transportation domain, and is one of the important parts in transportation management. In this thesis, a vehicle license plate is recognized and outputted in text form. The target is to recognize the character for information management and safe guarantee.In the license plate location, after researching on more than 200 vehicle pictures, we find that the license plate area can be outstood by differential operation of the saturation and mathematical morphology operation, then the accurate license plate position can be obtained by geometric feature. In the license plate normalization processing, the contrast of the image is enhanced by using gray image equalization. Employing Hough Transform and projection method; the slant plate is corrected.In the segmentation, utilizing adaptive multi-scale template matching method presented in this thesis, the characters are segmented. This method can reduce the impaction of character conglutination, character rupture and noise on the segmentation. In the character recognition, the characters are normalized and refined. Then employing improved Hausdorff distance template matching algorithm, the number and the letter are recognized finally.The results of experiments show that the algorithm has small amount of computation, strong anti-noise capability and satisfactory recognition rate. It can locate vehicle license plate, segment characters and recognize them accurately. The algorithm can be used on the modern traffic management.