Evaluating the Effectiveness of Supervised and Unsupervised Classification Methods in Monitoring Regs (Case Study: Jazmourian Reg)

Due to its mobility and ability to move and its direct impact on residential areas and various developmental activities, the Ergs are of major importance in the desert areas, so monitoring of those is very important. Considering that the use of supervised and unguarded methods is considered as one of the most common methods in determining and monitoring land uses, in this research, the accuracy of different classification methods in the monitoring of Jazmourian regs was investigated. In this research, in order to evaluate the accuracy of different classification methods in determining the type of usage of the study area, especially the regs, satellite image of the Landsat 8, OLI sensor in 2017 has been used. First, the area of ​​the regs prepared manually using Google Earth and Topography Map of the region, and then in the ENVI software, the land use type of ​​the study area was determined by different supervised (Maximum Probability, Minimum Distance to Mean, equilibrium, Parallelepiped) and unsupervised methods (K-mean) and then the plotted area using two point and surface methods were compared with the area determined by different classification methods, and the accuracy of each method was measured. Due to the similarity of the reflection type of Landsat images in desert areas, the results are of little precision, so that the evaluation results indicate that in the point and surface method, maximum probability classification, with a total accuracy of 64.9 % and 53% has the highest accuracy and the k-mean method with a total accuracy of 15.5% and 17% has the lowest accuracy, respectively. So, in order to monitor the type of land use, including desert areas regs, a different kind of satellite imagery or other classification algorithms should be used.