In order to describe image complexity in a better way

this paper discussed the description principle of traditional complexity from the perspective of composition theory

and introduced it to the domain of image complexity. Firstly

the authors discussed traditional calculation method of complexity and analyzed its limitation

which is based on a sequence of tests. Secondly

the authors proposed a novel approach to image complexity description

which adopts gray level

gray spatial distribution

and object quantity

including information entropy

texture

edge

and other relevant factors. Finally

the complexity of several different remote sensing images was tested. Results show that this description way can be used to reflect the overall image complexity

achieving the purpose of quantitative and quantitative comparisons between different images.