Doctoral defence: Kalev Koppel "Advancing Urban and Agricultural Monitoring Using Sentinel-1 Synthetic Aperture Radar Data"

On the 5th of May at 12:15, Kalev Koppel will defend his doctoral thesis, "Advancing Urban and Agricultural Monitoring Using Sentinel-1 Synthetic Aperture Radar Data", to obtain the degree of Doctor of Philosophy in geoinformatics.

Supervisors:
professor Tõnu Oja, University of Tartu

Opponent:
professor Ramon F. Hanssen, Delft University of Technology, The Netherlands

Summary:
The dissertation focuses on the use of Sentinel-1 synthetic aperture radar (SAR) data, provided by the European Space Agency, for the remote sensing of urban areas and grasslands. Compared to optical satellite imagery, radar images have the distinct advantage of being unaffected by cloud cover and sunlight conditions. This ensures consistent and frequent data availability, enabling the development of smart monitoring solutions.

To detect buildings from radar images, I developed a new method and compared its performance to previously established approaches. The study focused on two regions: Tallinn and Ida-Virumaa. The results were promising. My method achieved the highest accuracy in densely built-up areas, although its performance slightly lagged in regions with more sparse housing patterns.

Next, I investigated how building characteristics, such as height, material, orientation, and shape, influence radar signal backscatter. Several patterns, rooted in both practical reasoning and physical theory, were confirmed. For instance, larger and taller structures are more visible to radar, and metal reflects more strongly than wood. However, buildings positioned at certain angles may become nearly invisible to radar.

Finally, I studied how SAR is able to detect agricultural events in grasslands. To get subsidies the farmers must mow grasslands within a specific timeframe. Radar time-series data revealed distinct patterns for both mowing and ploughing events, each leaving a unique “fingerprint.” Using this insight, remote sensing company KappaZeta OÜ, in collaboration with Tõravere Observatory, CGI Eesti AS, and STACC OÜ, developed a mowing detection system for the paying agency.

In conclusion, my research demonstrated the significant potential of Sentinel-1 synthetic aperture radar data for urban and agricultural remote sensing. It also provided a foundation for the development of innovative applications in these domains.