吉林大学学报(信息科学版)

• 论文 • 上一篇    下一篇

图像去噪的混合滤波方法

项力领a, 刘智a, 齐冀b, 杨阳a   

  1. 长春理工大学 a. 电子信息工程学院; b. 光电工程学院, 长春 130022
  • 收稿日期:2013-03-13 出版日期:2013-05-27 发布日期:2013-06-07
  • 作者简介:项力领(1988—), 男, 湖北黄冈人, 长春理工大学硕士研究生, 主要从事图像处理研究, (Tel)86-15948034697(E-mail)469461223@qq.com; 刘智(1971—), 男, 河北定州人, 长春理工大学教授, 博士, 硕士生导师, 主要从事固态成像系统及其应用技术研究, (Tel)86-18946557983(E-mail)liuzhi@cust.edu.cn。
  • 基金资助:

    吉林省教育厅“十一五”科学技术研究基金资助项目(2010JYT04)

Method of Video Image De-Noising Based on Mixed Filter

XIANG Li-linga, LIU Zhia, QI Jib, YANG Yanga   

  1. a. College of Electronic and Information Technology; b. College of Photo-Electronic Engineering, Changchun University of Science and Technology, Changchun 130022, China
  • Received:2013-03-13 Online:2013-05-27 Published:2013-06-07

摘要:

针对视频图像在同时受到高斯噪声和脉冲噪声污染时,严重影响图像的存储、 编解码、 传输、 目标识别与跟踪的问题, 提出一种图像去噪的混合滤波方法。该方法通过基于个数判断脉冲噪声的方法, 将脉冲噪声从混合噪声中分离, 并利用中值滤波将其过滤; 再利用分块平均边缘检测的方法提取图像的边缘; 利用自适应均值滤波方法滤除非边缘的高斯噪声, 并将边缘图像嵌入滤除高斯噪声的图像中。实验结果表明, 该方法不但能有效去除图像中的高斯噪声和脉冲噪声, 而且能保持图像的边缘信息, 从而提高图像的去噪效果和清晰度。

关键词: 图像处理, 混合滤波, 个数判断脉冲噪声, 边缘检测, 高斯噪声, 脉冲噪声

Abstract:

Video image mixed noise by Gaussian noise and impulse noise, seriously affected the image storage, coding and decoding, transmission, target identification and tracking postprocessing. A mixed filtering method has been proposed based on an average of edge detection. Through the number of impulse noise judgment methods, the impulse noise is separated from the mixed noise and removed using median filtering. The edge of new image will be extracted by an average of edge detection method and then the Gaussian noise of nonedge is filtered using adaptive mean filtering methods. The edge of the new image is embedded in the filtered Gaussian noise images. Experiment results show that the method will be able to effectively remove Gaussian noise and impulse noise of the image, and to maintain the edges of the image information, improving the image de-noising and clarity.

Key words: image processing, mixed filter, the number of impulse noise judgment, edgedetecting, Gaussian noise, impulse noise

中图分类号: 

  • TP391.4