Posthurricane damage assessment using satellite imagery and geolocation features.

Area of Science:

• Disaster management and remote sensing.
• Geospatial analysis and computer vision.

Background:

• Timely situation awareness is critical for emergency managers and first responders post-hazard events.
• Satellite and drone imagery are increasingly used for damage assessment, specifically quantifying flooded or damaged buildings.

Purpose of the Study:

• To propose and evaluate a mixed-data approach for identifying damaged buildings after hurricanes.
• To improve the accuracy of damage assessment by integrating multiple data sources.

Main Methods:

• Leveraging publicly available satellite imagery and geolocation features of affected areas.
• Developing a mixed-data model combining imagery and geospatial data for hurricane damage assessment.
• Conducting a case study using Hurricane Harvey data in the Greater Houston area.

Main Results:

• The mixed-data approach demonstrated significant improvement over methods using only imagery features.
• The study successfully identified damaged buildings with enhanced accuracy.
• The findings highlight the benefit of incorporating geolocation data into remote sensing-based damage assessments.

Conclusions:

• Integrating geolocation features with satellite imagery offers a more robust method for hurricane damage assessment.
• This approach enhances the potential for unifying advancements in computer vision and traditional damage assessment methods.
• The openly available dataset facilitates further research in disaster response and remote sensing technologies.