The major difficulty for texture analysis in remote sensing is that no standard texture categories
exist since the visual effects of texture features depend on the spatial resolution of the image data
as well as the land cover characteristics of interest in specific applications [39]. In this study, to
reduce the searching space of similar texture patterns, each region is first assigned to one or more
categories in terms of the significant LCLU classes. Conceptually, suppose K is the number of