Identification of fuzzy objects from field observation data

This paper introduces the concept of fuzzy objects for modeling natural phenomena measured by field observation data. The propagation of uncertainties resulting from stochastic data errors and classification fuzziness is discussed, especially the interaction between these two kinds of uncertainties. Four object models are proposed to represent objects of different uncertainty levels. The proposed methodology is illustrated by a case study in coastal geomorphology of Ameland, The Netherlands. The methodology developed in this paper will also be applicable for modeling natural environments and physical processes in other fields.