RGB图像分形维数计算方法研究

STUDY ON COMPUTING METHOD OF FRACTAL DIMENSION OF RGB IMAGE

在计算颜色表面的分形维数时,为充分考虑颜色表面的相关信息,提出将RGB图像的3个分量R(红色)、G(绿色)、B(蓝色)重组为一个颜色矢量的思想,通过连接各像素点的矢量终端构建一个粗糙颜色表面,基于改进的投影覆盖法计算该表面的分形维数。与传统的灰度表面分形维数计算方法进行比较分析,分析结果揭示:与灰度表面维数计算方法相比,颜色表面维数计算方法通过以一个三维矩阵保存了RGB图像各像素点的R,G,B分量值,来更全面捕获真彩图像颜色信息,因而可以很好地克服灰度表面维数计算过程中出现的缺陷与不足——缺少对颜色的色彩和色饱和度的考虑。

In calculation fractal dimension of color surface,in order to adequately consider relative information of color surface,an idea that component of RGB images—R(red),G(green),B(blue) is recomposed into color vector is presented. By connecting vector endpoint of different pixels to construct a coarse color surface,based on modified projective covering method,the dimension of the surface is computed,and the comparison with traditional fractal dimension of gray-scale surface is made. The result discloses that comparative to computing method of gray surface dimension,by saving R,G,B component of different pixels of RGB image in 3D matrix,the computing method of color surface dimension catches more sufficiently color information of true-color image. Thus,it can better overcome the limitation and defect——lack of consideration towards hue and saturation during the course of dimension computation of gray surface.