图像平滑主要是为了消除噪声。噪声并不限于人眼所能看的见的失真和变形，有些噪声只有在进行图像处理时才可以发现。图像的常见噪声主要有加性噪声、乘性噪声和量化噪声等。图像中的噪声往往和信号交织在一起，尤其是乘性噪声，如果平滑不当，就会使图像本身的细节如边界轮廓、线条等变的模糊不清，如何既平滑掉噪声有尽量保持图像细节，是图像平滑主要研究的任务。-image smoothing is mainly to eliminate noise. Noise is not confined to the human eye can see, see the distortion and deformation, and some noise only when carrying out image processing can be found. The common image noise is the main additive noise, multiplicative noise and quantization noise. Image noise and signal often intertwined, especially multiplicative noise, if improperly smooth, images will make their own profile details such as borders, such as changing lines blurred, Smoothing out how to both noise as possible image details of the image smoothing main task.