Estimating Cultivation Area of Some Selected Crops in Bastam Plain by Using Multi-Temporal Sentinel-2 Images

The purpose of this study was to use the Multi-Temporal Sentinel-2 images and phenological index in separating and determining the cultivated area of the agricultural land in the Bastam region. To this end, the agricultural crops of the region were identified according to their types and phenological periods comprising apricot, grape, wheat, and forage corn. Three classifiers including support vector machine, maximum likelihood, and minimum distance models and field observations (points and boundaries provided by GPS) were used in order to compile a land use prediction map. Comparison of the accuracy of the three models showed that the support vector machine had the best performance, with overall accuracy and kappa coefficient of 0.86 and 0.82, respectively. The minimum distance model had the lowest classification performance with overall accuracy and kappa coefficient of 0.69 and 0.61, respectively. According to the model of support vector machine, the highest area (3423 hectares) was obtained for wheat, and the lowest was predicted for forage corn (738 hectares). Finally, the results showed that multi-temporal images and the phenological index had an acceptable capability for separation of the crops, prediction of their areas, and making suitable agricultural land use maps for the study area.