MSSC Initialization

This node performs the initial MSSC clustering based on the input data. The MSSC model (Mu and Wang 2008) follows the values of spatial order along the Peano curve with breaking points that are defined by a threshold population size. Through iterations in programming, each cluster satisfies the criteria of ascending spatial order and aggregation volume minimum constraints. MSSC (Mu and Wang 2008) is developed based on scale-space theory, an earlier algorithm, and applications of the theory in remote sensing and GIS. Using analogies of solid melting and viewing images, scale-space theory treats “scale”—corresponding to temperature in solid melting or distance in viewing images—as a parameter in describing the processes and phenomena. With the increase of scale (as temperature in the melting algorithm), a piece of metal will melt into liquid but not evenly, showing a clustering pattern; with the increase of scale (as distance in the blurring algorithm), the same image can reveal different levels of generalizations and details, or different cluster centers.