A Algorithm Used In Research To Segmentation And Localization For The Distorted License Plate

With the rapid development of information and the artificial intelligence technology, the technology of traffic management and control put forward higher request. In order to improve the correct recognition rate of license plate recognition is in the not ideal environment. In this paper, the deformation of the license plate location and license plate recognition key technologies are studied, specific to do the work as follow:The first, part of license plate locating. Given the license plate localization method based on canny edge detection.First, we use a way by canny edge detection operator to get license plate images, and then fitting a straight line to the edge image for getting the license plate candidate area of the image. To finally get to contain a license plate image connected analysis with precise positioning license plate candidate area.Secondly, part of license plate distortion correction. This article mainly based on perspective transformation is proposed distortion license plate correction algorithm. Aimed at the traditional license plate license plate image processing deformation correction algorithm of the deficiencies, we put forward a correction algorithm based on perspective transformation, then explain the correction, designing scheme and its application method in detail, this paper introduces principle of the algorithm, the process, steps, finally the result analysis.Thirdly, part of license plate character segmentation. In this paper, we show a kind of based on the improved vertical projection of license plate character segmentation method.First we making a preprocessing of license plate image after the location and correction before segmentation, such as remove frame and rivet, then normalized and projected characters in an image, analysis of adhesion which after projection. Finally, rearranging and normalized the results of segmentation.After repeated authentication and a lot of experimental analysis, this paper results show that the method used in this article to a serious distortion of the license plate image has good correction effect, and and to improve the recognition correct rate of corresponding license plate recognition system which under the undesirability conditions.