全变分耦合图像去噪模型

针对TV模型去噪后图像容易产生“阶梯效应”的现象，提出一种全变分耦合图像去噪模型。首先，根据去噪过程中图像梯度的变化趋势，构造一个趋势保真项，该保真项不但能有效去除图像噪声，而且能抑制“阶梯效应”。然后用小波在频域里对图像进行系数分解，利用Canny算法的边缘检测特性，设计控制函数，控制能量的扩散方向，保持了TV模型和趋势保真项的优点，能够在保护图像边缘纹理等细节信息的同时，抑制“阶梯效应”。实验结果表明，新模型的峰值信噪比、结构相似度、视觉效果均有显著提高。另外，所提模型的运行时间较短。

Coupling image denoising model based on total variation

The total variation (TV) model used in image denoising may produce “staircase effect”. A coupling image de-noising model based on total variation was proposed. First, a trend fidelity term based on the change tendency of image gradient was established. The fidelity term could not only remove image noise, but also restrain “staircase effect”. Then, wavelet was used to decompose coefficient in frequency domain, control based on the edge detection ability of Canny algorithm were designed. The control functions control energy spread direction, the advantages of TV model and trend fidelity term are maintained, edge and texture details were protected, and “staircase effect' was also suppressed. Experiment results show that peak signal to noise ratio (PSNR), structure similarity (SSIM) and visual effects of the nov-el model are much better. Moreover, the running time of the novel model is shorter.