基于CB模型的彩色图像混合噪声去除方法

摘要

将分数阶偏微分理论和CB模型相结合应用于图像去噪，提出了一种基于分数阶偏微分方程和CB模型的彩色图像混合噪声去除方法。对于添加混合噪声（高斯噪声和椒盐噪声）的彩色图像，首先，利用MCM模型去除图像中的椒盐噪声，然后将处理后的彩色图像分解为色度C和亮度B两部分，用分数阶偏微分模型处理亮度B，而对于色度C，由于其受到单位长度的限制，在处理时非常困难，利用拉格朗日乘数法并通过添加辅助变量，将色度转化为两个近似的子问题，从而得到色度的近似处理方法，最后将处理后的亮度B和色度C合成为新的彩色图像。最后通过实验证明了该方法的有效性。%Combining fractional-order differential theory with Chromaticity-Brightness(CB)model in order to deal with the mixture of the salt&pepper and Gaussian noises,a novel color image denoising model is proposed,which is based on fractional-order partial differential equation and CB model. For a color image with mixed noises(salt &pepper and Gaussian noises),the salt & pepper noise can be eliminated by the MCM model effectively. Then ,the processed color image is decomposed into chromaticity component and brightness component. Secondly ,we use fractional-order differential model for brightness component. For chromaticity component ,we use Lagrange multiplier method and add an auxiliary variable to approximate the chromaticity. Thirdly ,the retorted image is got by multiplying the recovered chromaticity with recovered brightness. Finally,we prove the validity of the proposed model through the experiments.