Spectral Mixture Analysis of the Urban Landscape in Indianapolis with Landsat ETM+ Imagery

• DOI: 10.14358/pers.70.9.1053
• Publication Date: 2013-11-27T02:30:17Z
• Authors: Dengsheng Lu; Qihao Weng
• Keywords: Thematic Mapper

This paper examines characteristics of urban land-use and land-cover (LULC) classes using spectral mixture analysis (SMA), develops a conceptual model for characterizing LULC patterns. A Landsat Enhanced Thematic Mapper Plus (ETM+) image Indianapolis City was used in this research minimum noise fraction (MNF) transform employed to convert the ETM+ into principal components. Five endmembers (shade, green vegetation, impervious surface, dry soil, dark soil) were selected, an unconstrained least-squares solution un-mix MNF components images. Different combinations three or four evaluated. The best images chosen classify based on hybrid procedure that combined maximum-likelihood decision-tree algorithms. results indicate SMAbased approach significantly improved classification accuracy as compared classifier. found be effective landscape