Bijan Younesi; Naser Ahmadi Sani; Soran Sharafi
Abstract
Agriculture is the basis for development and identification of crops and orchards is an important parameter in agricultural management helping planners through providing precise crop/orchard mapping. In order to overcome the limitations of fieldwork in crop and orchard mapping, satellite images seem ...
Read More
Agriculture is the basis for development and identification of crops and orchards is an important parameter in agricultural management helping planners through providing precise crop/orchard mapping. In order to overcome the limitations of fieldwork in crop and orchard mapping, satellite images seem to be appropriate due to providing wide coverage, timely and sequentially repeated image acquisition. In this study, IRS-P6 satellite images were analyzed in the Sharveran Plain lands in Mahabad County for orchard mapping. Various spectral indices were extracted using band ratioing and Principal Components Analysis (PCA) methods. Different supervised classifiers were used for classification of a 7-class (land use) and a 2-class (orchard and non-orchard) scenario. The classified maps were evaluated using the ground truth maps. The best overall accuracy and kappa coefficient were 97.95% and 0.95, respectively, using Maximum Liklihood classifier in the 2-class scenario. The results showed that IRS-P6 data are very suitable for identification and monitoring of orchards in terms of cost, time and accuracy.