经验模式多尺度图像分解

Empirical Mode Multi-Scale Image Decomposition

通常一幅图像包含有大量不同尺度的信息,而图像的应用研究往往仅限于某一尺度或某些尺度上的现象。因此,将图像按尺度进行分离十分必要,既可消除其他尺度信息对图像处理结果的影响,也可简化图像处理的难度和复杂性。笔者利用经验模式分解法,将图像按尺度从小到大进行分离。小尺度信息包含了图像的细节信息,剩余的大尺度信息表达了图像的基本趋势和结构。该方法为图像处理提供了一种有效的预处理方法。

An image generally includes a great deal of different-scale information, and that the image application research sometimes limits to the phenomenon of one scale or some scales. So, it is necessary to decompose an image according to the different-scale. It can eliminate the influence of other scales to the result of the image processing, and also simplify the difficulty and complexity of the image processing. This paper uses Empirical Mode Decomposition method to decompose image from small-scale to large-scale. The small-scale information contains the detail of the image, and the remained large-scale information expresses the prime tendency and structure of the image. This method is an efficient way in image preprocessing.