四种偏微分方程的图像去噪技术比较分析

Comparative Analysis and Research on Image Denoising Techniques of Four Partial Differential Equations

为了解决在去噪过程中图像的去噪与图像的边缘细节特征保持的矛盾问题,弱化传统的去噪方法在去噪时会破坏边缘、纹理等细节特征的影响,通过理论分析和MATLAB软件仿真实验相结合,对四种基于偏微分方程的图像去噪方法进行比较。利用图像质量评价标准S N对选择平滑扩散模型、各向异性扩散模型、全变差模型和四阶偏微分方程模型进行去噪效果评价。结果表明,四种偏微分方程的图像去噪技术均具有良好的去噪效果,按去噪效果由大到小排序为全变差模型、各向异性扩散模型、四阶偏微分方程模型、平滑扩散模型。

In order to solve the contradiction between image denoising and image edge detail feature preservation in the process of denoising,the influence of traditional denoising methods that will destroy edge,texture and other detail features during denoising is weakened.Through theoretical analysis combined with MATLAB software simulation experiments,four methods of image denoising based on partial differential equations are compared.The smooth diffusion model,anisotropic diffusion model,total variation model and fourth-order partial differential equation model were selected to evaluate the denoising effect using the image quality evaluation standard S N.The results show that the four partial differential equation image denoising techniques have good denoising effect.Sorted in descending order of denoising effect:total variation model,anisotropic diffusion model,fourth-order model,smoothing model.