Fuzzy Sets for Change Detection on Historic Forest Maps
要旨When studying forest cover change over the last 300 years, data from different sources and different positional accuracies are combined. A substantial amount of the detected change, based on the comparison of those maps, is false since it is caused by errors in the input material. Aggregation techniques, applied to reduce this error, may not be fit to reduce spurious change. This study presents a method to apply fuzzy sets and the epsilon band concept to locate real changes using historic forest maps. The result of this technique resembles the analysis of a human interpreter.