The comparison of fuzzy and maximum likelihood methods in preparing of land use layer using ETM+ data (case study: kameh watershed)

Land cover information can be used in hydrologic modeling to estimate the value of surface roughness or friction. The objective of this work is comparison of two different methods to provide land cover maps using Landsat images to estimate surface roughness in the manning equation and curve number (CN) in SCS method in the Kameh watershed located in the North of Torbat Heydarieh. At the first, Radiometric and Geometric correction performed on ETM+ data. Then with field surveying, The land cover classes were defined and training areas were selected. All of the bands with the exception of band 6 were used in classification. Because of low accuracy of village and road classes, These classes were removed from classification process and for entering these classes in the final map, information from the other layers of GIS system were applied. Results of this work show that in the fuzzy method using 3 layers in classification, The overall accuracy is 75.12% and kappa Index is 0.63. While those of fuzzy method using 2 layers in classification and maximum likelihood method are (73.3%, 0.6) and (72.39%, 0.59) respectively. Therefore the fuzzy method using 3 layers in classification is recommended.