小波变换用于图像处理之所以具有特别的优势，因为它能够聚焦到图像的细微变化。与小波变换不同，用偏微分方程来进行图像处理则需要对图像进行反复的迭代，直到得到一个稳定解，其整个过程是对图像进行整体处理，而且在去噪的同时，可以很好地保持边缘形状不变。若将两者有机结合，则不仅会克服各自的缺点，而且会得到更好的效果。为了在进行图像处理时，既能很好地抑制噪声，又能尽可能多地保持图像细节，提出了一种小波变换与非线性尺度扩散相结合的图像处理方法。该方法是利用小波变换的时频局部性和非线性尺度扩散的边缘增强特性来对图像进行处理，实验结果表明，其不仅能很好地抑制噪声，而且可保留尽可能多的图像细节，可见该方法是有效的。

Wavelet transform has its unique advantage in image processing, which analyze image subtly. Different from wavelet transform, computation using partial differential equations need to be iterated again and again. It deals with the image us a whole in the whole course. It can not only remove noise but also keep details of image. Combining with two methods, drawbacks will be overcome and better effect can be achiered. A nonlinear method for combining wavelet transform with nonlinear scale diffusion is proposed. By using the properties of time-frequency of wavelet transform and enhancing edges of nonlinear scale diffusion, the test results show the method is valid.