Multisource Data Integration for Georgia Land-Cover Mapping

Abstract

This paper describes a multisource data integration method for creating a 28-class 1998 land-cover map of the state of Georgia. The methodology for creating the map is unique in both the number and the method of integrating data sources. An assessment of the map's accuracy in the metropolitan Atlanta region using color infrared digital ortho-quarter quads and a preliminary state-wide assessment using low altitude aerial videography is presented. We estimate the overall accuracy for the metropolitan Atlanta region to be 86.3% and the overall accuracy for the state-wide land-cover map to be 83.9%. We anticipate that this dataset will be useful in a wide range of mapping, modeling, and planning exercises across the state and in the Southeast.