عنوان مقاله [English]
Background and Objectives: In Golestan province, the suitability of climatic condition to produce most of the agricultural products has led to high diversity in crop production, so this province has the first rank in terms of cultivating and producing oilseeds, especially soybean, in Iran. This research was carried out at four major watershed basins of Golestan province, Mohammad Abad, Qaresoo, Zaringol, and Gharnabad. This study was aimed to estimate the area under rice- and soybeans-cultivation in the aforementioned watershed basins. For this, Sentinel2 satellite images were used for the first time using different supervised classification methods (Maximum likelihood, the minimum distance of average and the Mahalanobis distance).
Materials and Methods:In this study, two Sentinel-2 satellite images of August and September of 2016 were used to identify, detect and evaluate the cultivated area of rice and soybean as two summer crops. This research was carried out at four watershed basins of Golestan Province (Mohammad Abad, Qaresoo, Zaringol, and Gharnabad). Radiometric, atmospheric, and geometric corrections were made after downloading the images of the study area. Then, band compounds, pseducolor combinations, image mosaics and rational band calculations were carried out, and the NDVI vegetation index was used to detect vegetation cover from other land uses, and finally, a land use map and crop layer was produced.
Results: results of this study showed that the soybean cultivation area which is an alternative plant for rice in summer cropping, has decreased compared to past years. In the present study, two Sentinel-2 satellite images of August and September of 2016 were used to identify, detect and evaluate the cultivated area of rice and soybean as two summer crops in four watershed basins of Golestan province. To compare the outputs of the three classification methods, training and test samples were used. In order to evaluate the accuracy of the classification results, the generated map was analyzed using the GPS-registered ground control point .The Maximum likelihood classification with kappa coefficient and overall accuracy of 92% and 95.5% was selected as the superior method for rice. In this method, the rice cultivation area was estimated 32911 hectares with a 18% bias compared to the Agricultural Jihad statistics (27839 hectares). Whereas for soybean, the minimum distance method with kappa coefficient and the overall accuracy of 88% and 95.2% was selected as superior classification method. Based on the results, the soybean cultivation area was estimated as 28359 hectares, with a bias of 13%, compared to the Agricultural Jihad statistics (25083 hectares).
Conclusion: Sentinel2 satellite images have a high potential for quick land detection and providing crops cultivation area maps in a regional scale. Also, the rice cultivation area has been increased compared to past years, while has been decreased for soybean.