Abstract: Principle and characteristic of P-M diffusion for image denoising are analyzed in theory，and the selection of parameter k of the P-M diffusion is analyzed also，the numerical calculation method of the gradient algorithm in the application of image denoising is discussed，and the improved P-M diffusion algorithm is obtained based on the method of traditional P-M diffusion.The results of experiment show that the improved algorithm can remove noise efficiently，and can keep the edges better than the traditional method of the P-M diffusion.

摘要： 从理论上分析了用P-M扩散方法去除图像噪声的原理与特点，分析了P-M扩散中参数k的选取问题，讨论了去噪应用中梯度算子的数值计算方法，在传统的P-M扩散方法的基础上得到了改进的P-M扩散算法。实验结果表明：与传统算法相比，改进的P-M扩散算法能有效抑制图像噪声，更好地保持边缘细节。