自适应扩散系数优化的图像降噪算法

An Adaptive Diffusion Coefficient Optimization Algorithm for Image Denoising

非线性扩散图像在降噪时纹理和细节通常会被削弱,为了解决这一问题,提出一种自适应扩散系数优化的图像降噪算法。利用经典的PM模型处理图像,利用扩散系数基于梯度幅值结合残差局部能量,可精确获取图像的纹理区域;利用绝对差值排序算子,进一步区分纹理部分及其存在的噪声;将图像梯度信息,残差能量及绝对差值排序算子融合到模型中,在去噪的同时很好地保留图像边缘、纹理等细节信息。实验结果表明,所提算法的SNR值为18.4714,RMSE值为15.8373,UQI值为0.8193,在降噪的同时较好地保留图像的纹理和细节,在视觉质量方面具有优越性能。

In order to solve this problem,an image denoising algorithm based on adaptive diffusion coefficient optimization is proposed.The classical PM model is used to process the image,and the diffusion coefficient based on gradient amplitude combined with residual local energy can accurately obtain the texture region of the image;The absolute difference sorting operator is used to further distinguish the texture part and its noise;The image gradient information,residual energy and absolute difference sorting operator are fused into the model to preserve the image edge,texture and other details while denoising.The experimental results show that the SNR value of the proposed algorithm is 18.4714,the RMSE value is 15.8373 and the uqi value is 0.8193.It can reduce noise while retaining the texture and detail of the image,and has superior performance in visual quality.