基于扩散张量的自适应正则化变分模型  

Adaptive regularized variation model based on diffusion tensor

作　　者：刘孝艳[1,2] 冯象初[1] 

机构地区：[1]西安电子科技大学理学院数学科学系,陕西西安710071 [2]西安石油大学理学院,陕西西安710065

出　　处：《系统工程与电子技术》2010年第1期188-191,194,共5页Systems Engineering and Electronics

基　　金：国家自然科学基金(60872138)资助课题

摘　　要：结构张量是描述图像的有效工具。利用结构张量对图像灰度变化的方向和大小进行判断,提出基于扩散张量的自适应正则化变分模型。该模型将冲击滤波器耦合在其中,使其在恢复图像的同时能有效地增强边缘。同时,给出一种构造正则化参数的方法。仿真实验表明,该模型在对带噪图像进行自适应恢复时,能较好地保护边缘信息,增强纹理特征,得到了较为满意的结果。Structure tensor is good for describing images, this paper puts forward an adaptive regularized variation model based on diffusion tensor. In this model, the direction of diffusion and the characters of different kinds of pixel in noisy images are characterized by the eigenvector and eigenvalues of structure tensor. In order to enhance edges, the shock filter is coupled to it. And the principle of selecting the parameters is presented. Simulation experiments indicate that the proposed model can not only denoise efficiently but also preserve detail information well, thus obtainining the better results.

关 键 词：图像恢复 扩散张量 自适应正则化 变分模型 冲击滤波器 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]