Detecting change in vague interpretations of landscapes

For a number of years researchers have advocated the use of fuzzy classifications in the study of land cover mapping from satellite imagery. Some studies have looked at the change of fuzzy spatial object, but none have considered the direct corollary of the so-called change detection matrix. In this paper we discuss populating the fuzzy change matrix, using fuzzy logic statements. Intersection is the principal operation, but it is argued that the Bounded Difference is the intersection operation for which the results make sense for determining loss and gain of a cover type. While the minimum operator works in actually populating the matrix. An alternative matrix can be generated using just the Bounded Difference. The detection of ecotones and the analysis of ecotone change are also discussed. It is suggested that the mappings derived express subtle variations in land cover types and change in those types as well as in ecotones, which may be related more conclusively to an ecological process than are Boolean mappings with associated linear boundaries.