摘要
文章通过分析调和去噪模型、ROF模型、广义TV模型,提出了基于TV模型的去噪算法改进。在调和模型的基础上对扩散系数进行了改进,从而克服了ROF模型的阶梯效应。改进后的算法可以去除更多的噪声,同时保留了图像的边缘,且与ROF模型相比,在不同噪声下其峰值信噪比均能有一定的提升。
This paper proposes an improvement of the denoising algorithm based on TV model by analyzing the harmonic denoising model, ROF model and generalized TV model. The diffusion coefficient is improved on the basis of the harmonic model, thus overcoming the ladder effect of the ROF model. The improved algorithm can remove more noise while preserving the edge of the image, and its peak signal-to-noise ratio can be improved under different noises compared with the ROF model.
引文
[1]RUDIN L I,OSHER S,FATEMI E.Nonlinear total variation based noise removal algorithms[J].Physica D Nonlinear Phenomena,1992(4):259-268.
[2]宋海英,荀月凤.全变分图像去噪的研究[J].成都电子机械高等专科学校学报,2009(4):23-25.
[3]许娟.图像去噪的ROF模型及仿真实验[J].安庆师范大学学报(自然科学版),2010(2):15-18.
[4]周先春,吴婷,石兰芳,等.一种基于曲率变分正则化的小波变换图像去噪方法[J].电子学报,2018(3):621-628.