| چکیده انگلیسی مقاله |
Jiroft Dam constructed on Halil River is the fifth concrete dam in Iran. Its reservoir was planned to irrigate more than 14200 ha of downstream agricultural lands including Jiroft case study. Due to population growth and increasing demand for agricultural products, this tropical region has experienced rapid landscape changes. This study aims to determine the land cover (LC) of Jiroft area using some supervised and unsupervised classification techniques, satellite images and geographic information system (GIS). In this regard, some pixel-based classification methods like mahalanobis distance (MD), maximum likelihood (ML), neural network algorithm (NN) and support vector machine (SVM) have been employed. Landsat 8 imagery data of OLI sensor for September, 2020 was acquired and its land cover was classified into five classes of orchard, agriculture, water body, rock and Barren lands. Finally, using ground control points (GCPs), derived by global positioning system (GPS), the performance of these classification methods were evaluated. Results showed kappa coefficient as well as overall accuracy for MD, ML, NN and SVM methods were equalling to (81%, 86%), (88%, 91%), (90%, 93%) and (88% and 92%), respectively. Comparison of results reveals a superior capability of NN method for land cover classification. |
