| With the development of information technologies, digital image processing is becoming to be a more and more important research subject. Supported by a solid theoretical foundation, digital image processing based on PDEs (partial differential equations) is used widely in related fields. Image denoising technology based on anisotropic diffusion equations is one of the digital image processing methods based on PDEs..Developed from traditional Gaussian filtering, it can well keep edge characteristics in the process of noise smoothing, and is widely recognized in image denoising apllications.Based on a theoretical study of anisotropic diffusion equations, this thesis mainly focuses on the application of anisotropic diffusion equations in image denoising., An improved P-M anisotropic diffusion model is proposed to enhance the application performances of the classical model. The experimental results showed that the propoesed model could effectively suppress speckle noise while keep edge characteristics in medical ultrasound images.On the other hand, with the development of information technologies, the traffic management system has been the tendency. License plate recognition (LPR) system is the core of the intelligent transportation system (ITS), and plays an important role in transport fare management systems. Researches on key technologies of LPR have become a hot issue in ITS.Some key problems in license plate location, license plate character segmentation and recognition are studied in this thesis. We mainly focused on the commonly used template matching and neural network based character recognition. In the method of template matching, some improvements are put forward and validated by experiments. In the method based on neural network, Zernike moment features of images are used in the character recognition, and good experiment results have been achieved. At the end of this article, comparative analyses have been made for template matching and neural network based methods. |
PDF Full Text Request