【作者】 李鹏；

【导师】 王俊平； 孙京梅；

【作者基本信息】 西安电子科技大学 ， 电子与通信工程， 2014， 硕士

【摘要】 数学形态学作为一门新兴的图像处理理论,凭借其坚实的理论基础和简单的算法,在图像处理各个领域均获得了广泛的应用。传统的数学形态学在图像处理时只采用一个形状与数值均不变的结构元素,如果结构元素选取的不恰当,往往会对处理结果造成不利的影响。本文对结构元素能够自适应选取的自适应形态学进行了研究,并在图像去噪和边缘检测应用中验证了它的优越性。本文首先对灰度图像领域的数学形态学做出研究,并分析了采用固定结构元素的传统形态学在图像去噪和边缘检测两个应用中存在的问题,随后针对这些问题,分别提出了能够根据输入图像局部特征而自适应选取结构元素的自适应形态学方法,并在仿真实验中分别对所提出方法的有效性进行了验证。实验证明所提出的自适应方法不仅较好地解决了传统形态学方法存在的问题,并且具有较好的处理结果。为了将所提出的针对灰度图像去噪应用的自适应形态学方法拓展到彩色图像领域,本文研究了了彩色图像的特性和几种不同的颜色矢量序,并在退化序和边际序两种不同的矢量序模式下分别对此方法进行了拓展。经实验证明,此方法在两种序下对彩色图像均具有较好的去噪效果。更多还原

【Abstract】 Mathematical morphology, as an emerging image processing theory, is widely used in various fields of image processing because of its solid theoretical foundation and simple algorithm. A structure element with fixed shape and value is always used by the traditional mathematical morphology, and it may leads to a bad influence to the processing result if it is inappropriate. The theory of adaptive mathematical morphology, which can select an appropriate structure element based on the local feature of an input image, is studied in this paper and some discussion is made to prove its advantages during the image denoising and edge-detection.The theory of gray mathematical morphology is firstly studied and the drawbacks of the traditional mathematical morphology which uses a fixed structure element in image denoising and edge detection are analyzed in this paper. Then two adaptive mathematical morphology methods which can select structure elements adaptively are proposed respectively to solve these drawbacks. The effectiveness of the proposed methods is proved through simulation experiments, and the results of the experiments show that the adaptive morphology methods not only solve the drawbacks of traditional morphology methods but also have good processing results.In order to extend the adaptive morphology method for gray image denoising to the color image field, the features of color image and some different color vector ordering methods are studied in this paper. Then the proposed adaptive morphology method for gray image denoising is respectively extended to the color image field through two vector ordering methods--reduced ordering method and marginal ordering method. Experiment results show that both of the proposed methods based on the two ordering methods have good denoising effect.更多还原