基于PCNN的彩色图像混合噪声滤除  

Mixed-noise Removal for Color Images Based on Pulse Coupled Neural Network

作　　者：何凯[1] 黎绍发[1] 王成[1] 

机构地区：[1]华南理工大学计算机科学与工程学院,广州510006

出　　处：《计算机工程》2010年第9期214-216,共3页Computer Engineering

基　　金：广东省工业攻关基金资助重点项目(2004B10101032)

摘　　要：针对基于阈值线性衰减输出带权均值型脉冲耦合神经网络(L&A-PCNN)的彩色图像混合噪声滤除在进行图像滤波时易产生噪点误判,从而导致污迹斑的问题,结合混合噪声性质,以及PCNN模型的点火特征,对L&A-PCNN算法做出进一步改进,建立新的噪点判断方法,提出基于L&A-PCNN的彩色图像混合噪声滤除改进型算法,仿真实验证明了其有效性。This paper proposes an improved mixed-noise removal method for color images based on Pulse Coupled Neural Network（PCNN） which has linear attenuated threshold and weighted averaged gray level output to solve the noise blotches problem with the estimation of noise pixels achieved by analyzing the features of mixed-noise and fire-time matrixes.Mixed-noise removals for various color images are implemented with the proposed algorithm and the experimental results demonstrate its validity.

关 键 词：脉冲耦合神经网络 混合噪声 阈值线性衰减 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]