Assessing Accuracy of Land-Cover Change Data Aggregated to a Fixed Spatial Support
Stephen V. Stehman, James D. Wickham
A method for assessing the accuracy of land-cover change products is described that evaluates net change at a specified spatial support. Accuracy is quantified by the mean absolute deviation, a metric derived from the absolute value of the difference between the map change in a land-cover class and the true change in the land-cover class. Several sampling design options are described for collecting the reference data. Stratification is a key design feature because of the desire to increase sample size in the rare high change areas of the map. A protocol for a priori evaluation of different sampling designs is described. Details of the accuracy assessment protocol are illustrated via a change product derived from the 1992 and preliminary 2001 National Land-Cover Data (NLCD).
Reference: Proceedings of the 7th International Conference on GeoComputation
, University of Southampton, United Kingdom, 8 - 10 September 2003.
CD-ROM. Produced by D.Martin, "GeoComputation CD-ROM".