一幅原始图像,在获取和传输过程中,会受到各种噪声的干扰,使图像退化,质量下降. 退化会引起图像模糊,特征淹没,对分析图像不利.为了抑制噪声改善图像质量进行的处理称为图像平滑或去噪.利用高斯函数得到了图像平滑去除随机噪声中使用的平滑模板的一个较普遍的公式。该公式为图像平滑去除噪声提供了多种模板选择。针对不同的图像噪声水平，选取不同的模板，才能达到去除噪声的最理想的效果。文中给出了图像平滑的实现算法,显示结果验证了上述方法和结论。-Any picture would be affected by all kinds of noise during the course of acquisition and transition which would lead to the degradation of a picture. The image and the character of a picture indistinct which is an disadvantage for image analysis. The way for lower down the noise to improve the quality of the picture calls the smoothing of image or denoising . It can be accomplished in spatial domain and frequency domain. A more common formula of templates used in image smooth denoising is developed from Gauss function in order to offer more templates to be chosen when denoising images with random noise. For each image with different levels of noise, only one template corresponding to the level of this image noise is optimal when denoising this image. The methods of smoothing processing are given in this paper .And the showing results are presented and they validate the above formula and conclusions.