摘要： 提出一种图像去噪中的有限元求解方法。从P-M模型出发，利用双线性插值得到图像在任意点的灰度值，将该值作为初始条件，采用有限元法对模型进行分析求解。在相同的时间间隔与迭代次数的条件下，该方法的求解结果与传统的有限差分法相比，平滑去噪和保持边界的效果更好，峰值信噪比有较大提高。

关键词: 非线性扩散方程, 有限元法, 有限差分法, 图像去噪

Abstract: This paper proposes a finite element solving method of image denoising. Based on the P-M model, it solves the image gray value at any point as initial condition by the bilinear interpolation method, then uses the Finite Element Method(FEM) to analyze and solve the model. Compared with traditional Finite Difference Method(FDM) at the same time intervals and iterations, it has better restoration effects of noise smoothing and maintain boundaries and greatly improve the peak value signal-to-noise ratio.

Key words: nonlinear diffusion equation, Finite Element Method(FEM), Finite Difference Method(FDM), image denoising

中图分类号: 

• TP391