Comparison of Maximum Likelihood and Minimum Distance to Mean Classifiers in Preparing Land Cover Map (A Case Study: Isfahan Area)

Land cover maps derived from satellite images play a key role in regional and national land cover assessments. In order to compare maximum likelihood and minimum distance to mean classifiers, LISS-III images from IRS-P6 satellite were acquired in August 2008 from the western part of Isfahan. First, the LISS-III image was georeferenced. The Root Mean Square error of less than one pixel was the result of registration. After creating false color composite and calculating transformed divergence index, the images were classified using maximum likelihood and minimum distance to mean classifiers into six categories including river, bare land, agricultural land, urban area, highway and rocky outcrops. The results of classification showed that the dominant land cover type is urban area, occupying about 6821.1 ha representing 38.86% of total area. The accuracy of maximum likelihood and minimum distance to mean classifiers was obtained using error matrix and Kappa analysis. According to the results, the maximum likelihood algorithm had an overall accuracy of 94.93% and the minimum distance to mean method was 85.25% accurate. The results illustrate that the maximum likelihood method is superior to minimum distance to mean classifier.