Research And Implementation Of License Plates Locating System Based On Edge Detection And Hough Transform

Automobile is one of important transportation tools in modern society, The car license has uniqueness as the automobile indication. Through gaining the number of car licenses, almost all information about vehicles would be clear. License Plate Automatic Recognition System is one of very important subsystems within IntelligentTransportation System, as the key to LPAR, the License Plates Locating system technology is of great significance.This paper summarizes the development and research results of the theory of image segmentation and License Plates Locating system, and analyzes the positioning methods used in the license plates locating system: firstly, uses the digital image processing techniques to pretreatment license plate images, and applies Canny operator for edge detection, Finally, applies Hough transformation to locate the rectangular region of license plate.Edge detection technology is foundation of such image processing technology as image measurement, image segmentation, image compression and pattern recognition technology, etc. How fast and accurate detection edge of image has always been the research hotspot in image processing. This paper focuses on an edge detection methods using Gaussian filter technology based Canny operator. Canny edge detection method has such advantages as accurately positioning, single-edge response and higher signal-to-noise ratio. This paper describes the basic ideology and specific methods of the algorithm, and large amounts of images experiments shows that Canny edge detection is an effective edge detection methods.It also analyzes Hough transformation. After detecting edge of images, the system uses Hough transform to locate parallel next borders and about borders of rectangular regional of plates.License Plates Locating system, which combines Canny edge detection methods and Hough transform, precisely locates the license plates region of license plate image in the experiments. Experimental results show that the target region of license plates images can be quickly and accurately detected.