Accuracy of landsat-TM and GIS rule-based methods for forest wetland classification in Maine

摘要: An investigation was undertaken to compare satellite image classification techniques delineate forest wetlands in Maine. Four were compared, including a GIS rule-based model. Accuracy assessments of the four methods on two study sites, Orono and Acadia, revealed very similar results. Overall accuracy for super groups (forest wetland, other upland, upland) ranged from 72% 81% at 74% 82% Acadia. Pairwise significance tests indicated that model significantly better than unsupervised both tasseled cap (Acadia) classifying groups. Although Kappa coefficients slightly higher compared hybrid classification, there no significant difference between either site. Forest wetland user's producer's 80% range highest achieved by or classification. Hydric soils, National Wetland Inventory data, slope percentage most important variables From this study, it appears combination are promising further investigations delineation.