快速随机值脉冲噪声的估计检测和滤波方法  

Fast methods for estimating,detecting and filtering random valued impulse noise level

作　　者：杨燕[1] 陈启强[1] 万毅[1] 

机构地区：[1]兰州大学信息科学与工程学院,兰州730000

出　　处：《兰州大学学报（自然科学版）》2010年第4期120-123,共4页Journal of Lanzhou University(Natural Sciences)

基　　金：教育部“新世纪优秀人才支持计划”项目(NCET-08-0258)

摘　　要：提出了一种快速、准确的随机值脉冲噪声水平的自动估计方法,并且对当前效果最好的基于对比度增强的滤波算法(CEF)进行了简化,使得噪声点检测的计算量大大减少.实验结果表明:该算法能够很好地自动估计噪声水平,同时简化了CEF算法,大大降低了CEF检测噪声点的计算复杂度,而且滤波效果仍能够达到甚至有时超过了CEF算法.A fast and accurate method was proposed for automatically estimating random valued impulse noise level. In addition, the current best denoising algorithm and the contrast enhancement-based filtering （CEF） algorithm were simplied by greatly reducing the computational cost during the noisy pixel detection stage. Simulation results confirm that our method can automatically estimate the noise level well and simplify the CEF algorithm, lowering its computational complexity significantly, with filtering performance equivalent to and sometimes even higher than the CEF algorithm.

关 键 词：随机值脉冲噪声 图像去噪 噪声水平估计 噪声点检测 

分 类 号：TN911.73[电子电信—通信与信息系统]