A Predictive Model to Assist the Public Sector in Structural Damage Categorization
Open AccessAccording to the National Centers for Environmental Information, there were twenty weather-driven disaster events exceeding $1B in the United States in 2021, resulting in 688 deaths and totaling an estimated $145B in damage. This Praxis develops methods to reduce time creating a traditional survey-based preliminary damage assessment by creating a predictive machine learning model. The basis of the model is persistent-scatter interferometric synthetic aperture radar derived from the European Space Agency’s Sentinel-1 mission at 5 x 20 meter resolution and ancillary geospatial information managed in a geographic information system. This research uses historical Federal Emergency Management Agency damage assessments from three hurricanes 2016 Matthew, 2017 Irma, and 2017 Harvey to develop and test the framework and subsequent models. This research builds on the current body of knowledge using machine learning with other remote sensing phenomenology and ancillary geospatial information to predict binary damage categorization within structures. The impact of this research is using persistent-scatter interferometric synthetic aperture radar data to create a damage categorization prediction model fused with geographic information system in hours vice weeks with traditional methods. The novel contributions of this research are developing a framework and rapid process to move from synthetic aperture radar single look complex images and structured geographic information system to categorical structural damage prediction at greater than 98% accuracy across all studied datasets.
- All rights reserved
Notice to Authors
If you are the author of this work and you have any questions about the information on this page, please use the Contact form to get in touch with us.