The Japanese companies Synspective and Spectee have developed a unique method for highly accurate estimation of the scale of floods, which combines satellite radar imagery and artificial intelligence to analyze social networks.
This hybrid approach allows you to eliminate the "blind spots" of both technologies based on the principle of merging "Eye in the sky" and "Eye on the ground" data.
How the integrated assessment system works:
Space monitoring (Synspective): The company uses its own constellation of StriX microsatellites. They are equipped with synthetic aperture radars (SAR), who can see through the clouds, showers and record flooding even at night. However, satellites are often "blinded" in dense urban development due to the blocking of the view by high-rise buildings.
Analytics on the ground (Spectee): Spectee uses AI to instantly collect and analyze text messages, photos and videos from the scene of the events, which are published by users in social networks. Algorithms recognize exact coordinates and even determine the current depth of street flooding in real time. The main disadvantage of this method is the complete lack of data in sparsely populated areas.
Data fusion result: Combining flood contours from space and point measurements from social networks, the system automatically fills in the gaps. Radar data helps to calculate river flooding in fields, and eyewitness posts - to clarify the situation at specific city intersections.
The effectiveness of the technology has already been officially confirmed during a retrospective analysis of severe floods in Kumamoto Prefecture. Flooding map, created by artificial intelligence, almost completely coincided with the data of real large-scale geodetic surveys, conducted by local authorities.
Expected, that Synspective and Spectee's joint solution will enable government agencies and insurance companies to assess losses instantly, carry out evacuation faster and coordinate rescuers effectively.
Source: https://spacemedia.jp