After obtaining classification results, it is necessary to check its accuracy. Assessment of the thematic accuracy
of the classification results is defined as the verification of the labeled pixel identity against the ground truth at the
time of sensing at representative sample points (Lillesand et al. 2004). Therefore, accuracy assessment has been
carried out and the kappa coefficient was calculated for all the three methods in order to select best classification
algorithm among them. During this process, around 40 ground truth points were chosen within the study area which
includes sample points in all LU/LC classes. Based on the results of accuracy assessment, best classification
algorithm was identified and its results were utilized for the process of change detection. LU/LC maps were
composed using ArcGIS 9.3 software and map to map comparison was used for change detection.