Evaluation of Sentinel-1 and Sentinel-2 Feature Sets for Delineating Agricultural Fields in Heterogeneous Landscapes

The Group on Earth Observations Global Agricultural Monitoring Initiative (GEOGLAM) considers agricultural fields as one of the essential variables that can be derived from satellite data. We evaluated accuracy at which delineated Sentinel-1 (S1) and Sentinel-2 (S2) images in different landscapes throughout growing season. used supervised segmentation based multiresolution (MRS) algorithm to first identify optimal feature set S1 S2 for field delineation. Based this set, we analyzed with increasing data availability between March October 2018. From sets, combination two polarizations radar indices attained best results. For S2, results were achieved using a all bands (coastal aerosol, water vapor, cirrus excluded) six spectral indices. Combining further improved Compared single-period dataset March, covering whole season led significant increase accuracy. very small (< 0.5 ha), obtained was 27.02%, (0.5 – 1.5 57.65%, medium (1.5 ha 15 75.71%, large (>15 stood 68.31%. As use case, result aggregate improve pixel-based crop type map Lower Saxony, Germany.