Integration of environmental variables with satellite images in regional scale vegetation classification

The difficulty of collecting information at conventional field studies and relatively coarse spatial and spectral resolution of Landsat images forced the use of environmental variables as ancillary data in vegetation mapping. The aim of this study is to increase the accuracy of species level vegetation classification incorporating environmental variables in the Amanos Mountains region of southern central Turkey. In the first part of the study, ordinary vegetation classification is attained by using a maximum likelihood method to Landsat images with the help of forest management maps. Discriminant analysis is used in the second part of the study in two different stages. First, Fisher's linear equations for each of the pre-defined nine vegetation groups are calculated and values of the pixels are assigned by the greatest probability value. Second, distance raster value of maximum likelihood classification is used as a threshold. The distance raster pixels having values less than one is accepted as misclassified and replaced with the results of discriminant analysis results. As a result of this study 19.6% increase in overall accuracy is obtained by using the relationships between environmental variables and vegetation existence.