一种图像分形特征提取的近似算法

An approximate algorithm for extracting fractal features of images

提出了一种提取图像分形特征的近似算法．该算法对经典的“毯覆盖”算法中按不同几何度量尺度分别计算图像的“膨胀层”和“腐蚀层”进而计算不同几何度量尺度图像表面积的计算过程进行了简化，降低了利用“毯覆盖”算法提取图像分形特征的计算量和存储量，在现有硬件条件下使实时提取图像分形特征成为可能．

An approximate algorithm for extracting fractal features of an image is presented. In the algorithm, the conventional Blanket covering algorithms computing procedure, in which the values of the surface area of an image in different measuring scales are obtained by computing the images dilating surfaces and eroding surfaces respectively in different measuring scales, is simplified, and the amount of computation and storage for extracting fractal features of an image by the Blanket covering algorithm is reduced. The algorithm makes it possible that the fractal features of an image could be extracted in real time in the existing hardware environment.