Fuzzy models for handling uncertainty in the integration of high resolution remotely sensed data and GIS

Abstract

The advent of new high resolution sensors, either airborne or spaceborne, leads to new applications and further impulses for an integration of remotely sensed and GIS data. Along with these new data sources, existing processing methods have to be adopted which is in particular also valid for the assessment of the post classification quality. In this overall context our contribution will outline general uncertainty aspects in the integration process and particular problems with the accuracy assessment based on high resolution data. These problems lead to the motivation to develop a new characteristic value, called the Fuzzy Certainty Measure (FCM), which considers indeterminate boundaries in the classification result as well as in the reference data, and can be applied in a class- and even object-specific manner. © 2007 Springer.