Image Segmentation And Its Application To Vehicle License Plate Identification

As an elementary and important part of digital image processing technique, image segmentation bears a broad prospect of applications and researches. From the viewpoint of image techniques, image segmentation is between image processing and image analysis. It is not like a common image processing such as image acquisition, image transforming, image enhancement and so on. And it is a basically important stage of image analysis. Because of its obvious importance, it has become one hotspot of research in the domain of image techniques.This thesis was mainly involved in the research of image segmentation, and review and sum up the work of my graduation. The main contributions include:1. Review the development of researches of algorithms of image segmentation and quality evaluation of image segmentation and illustratethe principle and extension of the classical methods--Otsu andmaximum entropy in details.2. Put forward a new algorithm--two-dimensional Otsu thresholdingsegmentation based on verdict scope adaptive restriction (VSAR). The new algorithm uses an experienced formula which adaptively determines the verdict scope. It not only reduces the running time, but also bears practicability in some applications.3. Bring forward a new algorithm--two-dimensional maximum entropythresholding segmentation based on verdict scope adaptive restriction (VSAR). It utilizes an experienced formula which adaptively determines the verdict scope. It not only reduces the running time, but also bears practicability in some applications.4. In order to locate the vehicle license plate in the research of vehicle license plate identification, a simple and useful method of locating vehicle license plate (VLP) called counting changing times of edge is proposed, which bears good performance in the simulation of application.