A simple polygon-polygon basis method to compare categorical raster maps

Abstract

In this paper we present a method for comparing categorical raster maps based on comparing categories at the polygon level. In contrast with the most commonly used pixel-level methods of comparison, the proposed method reduces both possible map registration errors and errors resulting from the uncertainty associated with defining the boundaries of each category. The results are presented visually as a map which shows the degree of fit between the compared maps, and analytically as an overall agreement index varying between 0 and 1 and by a fuzzy kappa index. The methodology is illustrated by two examples: the first compares two synthetic maps which were generated specifically to demonstrate the validity of the method, and the second example analyses the use of the method to detect changes over time in two real land-use maps which were made in different years.

Keywords

• raster model
• map comparison
• categorical data
• fuzzy kappa