Modeling Urban Landscape Dynamics Using Subpixel Fractions and Fuzzy Cellular Automata

Abstract

This study proposes a fuzzy cellular automata model based on the subpixel fractions extracted from multitemporal satellite images and discusses the relationship between sophisticated remote sensing techniques and an urban process model within the socioeconomic dimension. Accordingly, the major objectives of the present research are: (1) to incorporate the subpixel membership derived from remote sensing images into a fuzzy cellular automata model to simulate urban landscape change; (2) to standardize the quantitative method to incorporate subpixel information into a subcell cellular automata model and to find a better way to determine the parameters in the model's development, calibration, and validation. The comparison between the traditional cell-based model and subcell model suggests that the subpixel technique improves the accuracy of both urban mapping and modeling using medium resolution satellite images.