Vehicle License-Plate Recognition Based On Wavelet Transform And Vertical Edge Matching

Pattern recognition is one of the most important fields of the artificial intelligence. With the development of the artificial intelligence, the research on pattern recognition has been noticed by more and more people. Vehicle license-plate recognition is an actual application of pattern recognition.Vehicle License-Plate Recognition is an important tache in Intelligence Transportation System. It has a very important status in transportation surveillance and control, and it has been adopted widely to various applications, such as, toll-collection, managing parking facilities, and observation of stolen vehicles and monitoring of unauthorized vehicles entering private areas. Meanwhile, it also can be applied in detection and recognition fields. So, the technology is an important part in pattern recognition. The vehicle license-plate system can improve management efficiency, save money and manpower, and decrease commit rate about vehicles.License-plate locating, preprocess of image, binary of image, character segmentation and character recognition are mainly expatiated in this paper. By researching and summarizing these methods, and considering its specialty, some improved methods are put forward here.The precision of vehicle license-plate recognition lies on that of locating for the whole system. The locating of license-plate is mainly researched in this paper. Preprocess of image and edge detection are finished by using wavelet transform. Size-and-shape filtering on seed filling algorithm is provided. By using it, those noisy regions are removed. Then a vertical edge matching based algorithm recognizes license-plate from input gray-scale image.A new character recognition algorithm is presented, which combines the method using topology characteristics of characters and the method of artificial neural network. High recognition accuracy is reached.