针对图像脉冲噪声的参数自调整滤波算法(英文)  

New Efficient Automatic Parameter Tuning Approach for Removal of Impulse Noise

作　　者：陈大力[1] 薛定宇[1] 高道祥[1] 

机构地区：[1]东北大学人工智能与机器人研究所,沈阳110004

出　　处：《系统仿真学报》2007年第2期376-379,共4页Journal of System Simulation

摘　　要：提出了一种针对脉冲噪声的参数自调整图像滤波算法,该算法将基于相关性双阈值噪声检测的非线性滤波算法和基于最小相邻图像均方差的参数自调整算法有机地结合在一起,不需要了解原图像和噪声污染的信息,可以直接通过参数自调整算法对非线性滤波算法中的参数自动地进行优化选择。基于Matlab的仿真试验表明,该算法对脉冲噪声有非常好的抑制能力,并且能够很好地保护图像的细节信息,对各类不同密度的脉冲噪声图像进行滤波均能得到令人满意的结果。A new efficient approach to the restoration of images corrupted by impulse noise was presented. The proposed method combined a double-threshold nonlinear algorithm and a technique for automatic parameter tuning based on noise estimation. As a key feature, the method did not require any a priori knowledge about the density of noise corruption. Simulation results based on Matlab show that the filtering performance of the proposed approach is very satisfactory.

关 键 词：脉冲噪声 图像恢复 非线性滤波 噪声估计 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]