Insight Moment on the topic of “Automated Characterization of Exposed Buildings with Street-Level Imagery and Deep Learning", as a methodology of Exposure Modelling. 

Knowledge of the structural properties of buildings exposed to natural hazards is critical for accurate risk modelling and defining disaster management strategies. To date, however, such information is often unavailable or outdated.

Geotagged street-level imagery from global initiatives such as Google Street View yields a high potential for the automated inference of vulnerability-related building characteristics.
Learn more about it in this Insight Moment held by Mr. Aravena from the German Aerospace Center, in which he emphasizes the capability of the proposed approach by sharing experimental results for the Earthquake-prone Chilean capital Santiago de Chile. 

Patrick Aravena Pelizari received his Master of Science degree in Physical Geography/Environmental Systems from Ludwig-Maximilians-University Munich in 2013. Since 2013, he has been a research scientist with the German Remote Sensing Data Center (DFD) of the German Aerospace Center (DLR) in Weßling, Germany, where he has been contributing to several research projects in the fields of humanitarian relief and natural disaster management.

His work has focused on the development of machine learning methods for extracting related spatial information from multi-sensor remote sensing data as well as subsequent geostatistical analyses. His current research is on the development of deep learning methods for assessing building vulnerability in the context of natural hazard risk assessments, integrating Earth observation and geospatial data from multiple sources.


Join us for this Insight moment of the WG-RIUD on May 17th from 14:30 to 15:30h (CET time).