Hierarchical fuzzy pattern matching for the regional comparison of land use maps

Abstract

The evaluation of the spatial similarities between two raster maps is traditionally based on pixel-by-pixel comparison techniques. These procedures determine the number of cells in agreement for each landuse category and express the overall agreement with a boolean global similarity value. The problem with a pixel-by-pixel comparison is that a small displacement in pixels will be registered as disagreement even though the land use patterns between the maps maybe essentially the same. The issues of unique polygons mapping and hierarchical fuzzy pattern matching, where the maps are compared on both a local and global level emerge as viable and robust alternatives. The local matchings determine the degree of containment of each unique polygon to its spatial counterparts on the original maps in terms of fuzzy areal intersections. Local agreement values for the unique polygons based on their polygon area containments are calculated from fuzzy logical Max-Min compositional algorithms. A global agreement value is derived by the fuzzy summation of the local matchings. The uses of these basic methods are discussed and further refinements and modeling possibilities are outlined.