Investigation and Comparison of the Maximum Likelihood and Fuzzy Artmap Methods in Preparing and Monitoring Land Use Changes(Case Study: Omidiyeh, Khuzestan Province)

Remote sensing is the main technology to assess the expansion and rate of land cover changes. Knowledge of these changes in different parts has particular importance as the base information for different planning. The aim of the present study was the comparison of the maximum likelihood and fuzzy Artmap methods to prepare land use map using Landsat satellite images. In this study, changes in land cover were evaluated over the past 24 years in Omidiyeh. Sensor images of TM Landsat 4, ETM+ Landsat 7 and OLI Landsat 8 for the years 1990, 2000 and 2014, respectively, and topographic and land cover maps of the area also were used. Images of all three periods were classified into four land uses including rangelands, agriculture, saline lands, and residential areas. The overall accuracy of the classification results showed that the fuzzy classification method with a kappa coefficient of 93% had more accuracy in comparison with the maximum likelihood algorithm with a kappa coefficient of 81%. According to the obtained results, the most dynamic land use in the region was the agricultural area, whose extent increased from 1990 to 2014, and 32703.32 hectares (32.2%) was added to this land use. An increasing trend was also obtained for the residential land use with an area about 1325.1 hectares (0.94%). On the other hand, rangelands (13.4% of the study area) showed the highest reduction in area (18857.63 hectares).