INVESTIGATING THE POSSIBILITY OF PREPARING SMALL SCALE SOIL MOISTURE MAP FROM COUPLED SENTINEL-1 AND SENTINEL-2 DATA
Keywords: Soil Moisture Map, Object-Based Image Analysis, Multi-resolution Segmentation, Sentinel-1, Sentinel-2, SMAP
Abstract. With the failure of the radar instrument on NASA's Soil Moisture Active Passive (SMAP) satellite, the Sentinel-1 sensor has been considered as an alternative for replacing the SMAP radar data and restoring the combined radar and radiometer SMC product. A challenging subject to this purpose is the immense discrepancy between the spatial resolution of planned SMAP radar instrument (3 km) and Sentinel-1 data (10 m). In this paper, we investigate the possibility of preparing small scale soil moisture map and its quality from the synergy of Sentinel-1 and Sentinel-2 data using object-based image analysis (OBIA). To reach this goal, the most related features with soil moisture variable extracted from Sentinel-1 and Sentinel-2 data have been used as input layers to multi-resolution segmentation (MRS) algorithm to create image objects. Then the support vector regression (SVR) estimator has been used to calculate the soil moisture value of image objects. Initial evaluations demonstrate that produced soil moisture map obtained acceptable accuracy. In addition, the flexibility of the final product improves on the scale of the soil moisture map regarding the shape and size of image objects. It is also possible to combine this soil moisture product with a Level-3 SMAP SSM product to exploit the advantages of both products. This combination would lead to a small scale soil moisture map with enhanced accuracy and flexible scale.