Analysis of the complexity of remote sensing image and its role on image classification

A remote sensing image reflects the geographical environment of some region at a certain moment. The principal factor influencing the extraction of remote sensing information is the complexity of the image. This paper clarifies the complexity of a remote sensing image from the spectral domain, spatial domain and multi-spatial scale. The authors develop an expression for image complexity by quantitative analysis instead of an intuitionistic qualitative estimate. A quantitative index describing the complexity of the image information is presented based on the comprehensive description of the statistical property, Fourier transformation and information entropy of the image. Finally, a multi-level remote sensing information model is constructed based on different image complexities.