This research focused on conducting a crop-type study in Tehran Province, Iran, with a particular emphasis on wheat and barley, essential global agricultural products. Accurate mapping of these crops using remote sensing technologies is crucial for efficient agricultural management and planning. This study covered extensive areas within Tehran Province, including Rey, Varamin, Pakdasht, Pishva, and Qarchak.

A crop type map was created for wheat and barley crops, along with other agricultural products and non-agricultural areas, based on their phenological behavior using the agricultural calendar. Satellite images from Sentinel-1 and Sentinel-2 were used at key stages of crop growth, and features like NDVI, EVI, and VV/VH ratio were extracted to identify plant phenological trends. This study utilized the Google Earth Engine for efficient processing due to the large study area and volume of images.

Different scenarios were tested using a Random Forest classification algorithm with limited training data, resulting in the creation of a crop map. Scenario one, including various spectral bands and indices, achieved an accuracy of 84%, a Kappa coefficient of 69%, and an F1-score of 76%. Scenario two, focusing on spectral indices and the VV/VH ratio, obtained an accuracy of 87%, a Kappa coefficient of 62%, and an F1-score of 64%. The highest accuracy of 94%, Kappa coefficient of 87%, and F1-score of 88% were attained in scenario three by utilizing multispectral bands and VV/VH bands. Scenario four, using only spectral indices, achieved an accuracy of 73%. The superior performance of scenario three was credited to its comprehensive spectral and temporal information, demonstrating the effectiveness of remote sensing in large-scale agricultural mapping.

This research demonstrates the practicality and utility of using remote sensing for agricultural mapping in large areas. The methodologies and results of this research can significantly contribute to efficient monitoring and management of agricultural resources.