基于随机微分的EMD图像去噪算法  

An EMD algorithm for image denoising based on stochastic differential

作　　者：贺静波[1,2] 彭复员[1] 

机构地区：[1]华中科技大学电子与信息工程系,武汉430074 [2]海军工程大学电子工程学院,武汉430033

出　　处：《高技术通讯》2010年第10期1046-1048,共3页Chinese High Technology Letters

基　　金：国家自然科学基金(60872076)资助项目

摘　　要：考虑到传统的图像滤波算法在图像去噪的同时削弱了图像特征,以及图像系统所固有的自相似性和经验模式分解(EMD)算法的完备性和稳定性,提出了一种基于随机微分的改进EMD图像去噪算法。该算法首先利用EMD对图像进行分解,得到图像的多个固有模式函数(IMF)图像和剩余函数图像,然后根据IMF图像和剩余函数图像采取不同的随机微分滤波策略分别得到各层滤波结果,最后重组得到原始图像去噪后的结果。Matlab仿真证明,该算法在图像去噪的同时保留了图像特征。In consideration of the destruction to image features caused when the conventional Gaussian filter is used for image denoising, and image systems＇ intrinsic quality of self-comparability and the empirical mode decomposition （EMD） algorithm＇ completeness and stability, an EMD algorithm for image denoising based on stochastic differentiation is presented. The algorithm uses the EMD to decompose an image to obtain its numbers of intrinsic mode function （IMF） images and residual function images, and then uses different stochastic differential filters to obtain each layer＇s filtering result, and finally, the denoised image is given by rebuilding the IMF images and residual images. The simulation experiments are made by Matlab and the effectiveness of the algorithm is tested and verified.

关 键 词：经验模式分解(EMD) 随机微分(SDE) 噪声 福克尔-普朗克方程 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]