Classification of Coffee-Forest Landscapes Using Landsat TM Imagery and Spectral Mixture Analysis

• DOI: 10.14358/pers.79.5.457
• Publication Date: 2013-10-26T01:07:06Z
• Authors: Mikaela Schmitt-H...; Sean P. Sweeney; Tom P. Evans
• Keywords: Spectral Analysis; Thematic Mapper

This research applies linear spectral mixture analysis (LSMA) to a Landsat TM image, and assesses the value of fraction images (green vegetation, shade, soil) thermal band (TM-B6) in discriminating shade-grown coffee systems from forests. Four combinations bands were compared, maximum likelihood algorithm was used classify five land cover classes: high-density woodlands, low-density agroforests, crop / pasturelands, urban settlements. The classification accuracy each model combination assessed using both Kappa analyses quality allocation disagreement parameters. Results indicate improvements accuracies following inclusion TM-B6 as inputs classification; however, only use led significant at 95 percent confidence level. highest achieved 86 (Kstandard = 0.82), with producer’s user’s agroforests reaching 89 90 percent, respectively, an improvement over previous aimed spectrally distinguishing other woody types.