一种新的图像熵及其应用

In this paper, we induct the Boltzmann principle to the field of digital signal processing, construct a statistical theory of series, put forward the concept of Boltzmann entropy of image. This Boltzmann entropy compared with the Shannon entropy of image, we come to conclusions: the Boltzmann entropy of image reflects the spatial distribution characteristics of pixels in the nature, and could be used to measure the irregularity of the spatial distribution of the set of points. Finally, we make initial exploration with this new image entropy in image classification、image evaluation、image measurement and other fields, found many advantages with Shannon entropy.