Use of HCMM thermal data to improve accuracy of MSS land-surface classification mapping

Abstract

Abstract The use of thermal data from the Heat Capacity Mapping Mission (HCMM) satellite for delineating urban heat islands, and to aid in separating general land-cover types, was investigated. Day-time thermal infrared HCM M data were digitally integrated with LANDSAT Multispectral Scanner (MSS) data utilizing three different techniques and the combined data sets were subjected to unsupervised classification. Results from a comprehensive accuracy assessment show that the classification done by the procedurally least-complex merging technique yielded the highest overall accuracy and was consistently best for all cover-type categories and USGS 7·5′ topographic quadrangles evaluated in the study. Thus, even relatively coarse resolution (500 m) thermal data show promise for improving general land-surface classification accuracies over a large and diverse metropolitan region with man-made and natural cover types.