吉林大学学报(理学版)

• 计算机科学 • 上一篇    下一篇

基于全变分自适应保真项去噪算法的数值实现

王宏志 1, 刘婉军 1, 韩啸2   

  1. 1. 长春工业大学 计算机科学与工程学院, 长春 130012; 2. 吉林大学 学报编辑部, 长春 130012
  • 收稿日期:2014-03-10 出版日期:2014-11-26 发布日期:2014-12-11
  • 通讯作者: 韩啸 E-mail:hanxiao@jlu.edu.cn

Numerical Implementation of Adaptive Fidelity TermDenoising Algorithm Based on Total Variation

WANG Hongzhi 1, LIU Wanjun 1, HAN Xiao2   

  1. 1. College of Computer Science and Engineering, Changchun University of Technology, Changchun 130012, China;\=2. Editorial Department of Journal of Jilin University, Changchun 130012, China
  • Received:2014-03-10 Online:2014-11-26 Published:2014-12-11
  • Contact: hanxiao E-mail:hanxiao@jlu.edu.cn

摘要:

基于全变差图像去噪经典算法, 提出一种自适应保真项的数值实现算法. 该算法利用图像纹理区和光滑区中噪声的不同特点, 采用不同去噪强度避免传统方法的不足, 并以数值方法实现. 在保持经典算法去噪效果的前提下, 解决了原有阶梯效应和过度平滑的问题, 尤其对精致的纹理和细节图像, 使其在去噪的同时, 不丢失图像特点. 该方法处理相对简单, 可应用于以偏微分方程为基础的图像处理.

关键词: 全变分, 梯度抑制, 自适应, 纹理保持

Abstract:

On the basis of the classical algorithm of the image denoising based on total variation, a numerical algorithm of total variation based on adaptive fidelity term was proposed. Different intensities of denoising were used to avoid the shortages of the traditional method and then numerical method was chosen so as to realize our algorithm. On the premise of the classical total variation, our method made up for the shortcomings of the original ladder and excessive smoothing effect, especially for the image denoising of fine texture and detail images, it made the remain of most of their image characteristics. Our
 treatment can be applied to a series of image processing based on partial differential equations simply.

Key words: total variation, gradient inhibition, adaptive, texture preserving

中图分类号: 

  • TP181