Sa. Sader et al., ACCURACY OF LANDSAT-TM AND GIS RULE-BASED METHODS FOR FOREST WETLAND CLASSIFICATION IN MAINE, Remote sensing of environment, 53(3), 1995, pp. 133-144
An investigation was undertaken to compare satellite image classificat
ion techniques to delineate forest wetlands in Maine. Four classificat
ion techniques were compared, including a GIS rule-based model. Accura
cy assessments of the four methods on two study sites, Orono and Acadi
a, revealed very similar results. Overall accuracy for four super grou
ps (forest wetland, other wetland, forest upland, other upland) ranged
from 72% to 81% at Orono and 74% to 82% at Acadia. Pairwise significa
nce tests indicated that the GIS model was significantly better than u
nsupervised classification at both study sites, and significantly bett
er than tasseled cap (Acadia) in classifying the four super groups. Al
though Kappa coefficients were slightly higher for the GIS model compa
red to hybrid classification, there was no significant difference betw
een the two methods at either study site. Forest wetland user's and pr
oducer's accuracy was in the 80% range for the highest accuracy achiev
ed either by the GIS model or hybrid classification. Hydric soils, Nat
ional Wetland Inventory data, and slope percentage were the most impor
tant variables in the GIS model. From this study, it appears that a co
mbination of hybrid and GIS rule-based classification methods are the
most promising for further investigations of forest wetland delineatio
n.