一种改进偏微分方程扩散系数的图像去噪方法

An Image Denoising Method to Improve the Diffusion Coefficient of Partial Differential Equations

在偏微分方程P-M模型图像去噪过程中,扩散系数的选择会影响图像去噪的效果,为此提出了一个新的扩散系数模型来实现图像去噪。首先分析讨论了P-M模型中扩散系数和梯度阈值的选取对图像去噪的重要性,并对比了两个扩散系数的优点和缺点;在此基础上提出一个新的扩散系数,并应用到CLMC模型进行数值离散实验。实验结果表明,采用新的扩散系数得到的信噪比和峰值信噪比要比P-M方程中所给出的两个扩散系数得到的信噪比和峰值信噪比好。提出的扩散系数能够有效地进行图像去噪。

In the process of partial differential equations PM model image denoising, the diffusion coefficient will affect the image denoising effect is selected, have come out with a new model of the diffusion coefficient to achieve image denoising. First, discussed the importance of the diffusion coefficient of the PM model and the gradi- ent threshold for image denoising, and compare the advantages and disadvantages of the two diffusion coefficients, a new diffusion coefficient on this basis is proposed, and applied to the CLMC model discrete numerical experiments. Experimental results show that the new diffusion coefficients better than two diffusion coefficients of P-M model in PSNR and SNR. In this paper, the diffusion coefficient can be effectively denoising.