Downscaling Remotely Sensed Soil Moisture Variations Using Satellite And Unmanned Aerial Vehicle (UAV) Photogrammetry

Soil moisture is a critical component of the Earth's climate systems, impacting the coupled moisture-energy environmental cycles at the atmosphere-ground surface interface. Accurate estimation of soil moisture spatial variations is crucial for solving geotechnical engineering challenges (earth slopes and earth dam issues). From the first dedicated soil moisture satellite missions (SMOS, SMAP) to current missions such as Landsat 8 and Sentinel 3, the field of remote sensing of soil moisture has expanded. Satellite sensors are capable of observing broad areas, but their spatial resolution is limited by remote sensing data (microwave frequency, antenna diameters, altitude above the Earth's surface, and so on). The spatial resolution of the majority of passive radiometers is insufficient for estimating soil moisture. Additionally, recent improvements in remote sensing techniques and the use of Unmanned Aerial Vehicle (UAV) have demonstrated that soil moisture may be monitored using a number of distant sensing techniques. The fundamental objective of this work is to develop a soil moisture model employing an UAV and data analytics methods based on deep learning. Since August 2018, six roadway slopes in Jackson, Mississippi, have been comprehensively instrumented to record changes in moisture, matric suction, and temperature. The suggested model is based on over 15000 aerial photos taken by UAVs in a variety of weather and illumination situations. The surface soil moisture was connected with the UAV photos using remote sensing data processing, which was further validated using sensor data. The purpose of this study is to use intermediate UAV photos to downscale satelite-based soil moisture measurements to higher results.