At the time of flooding in many parts of China, DAMO Academy, Alibaba's research and innovation institute, urgently upgraded its remote sensing AI technology and developed water body recognition algorithms for flood control to support the Ministry of Water Resources' monitoring and analysis.
The number of image processing can be increased by 5 times than usual and the image analysis speed can be increased by 100 times in the key areas of over-alert water level.
In this flooding situation, due to climatic constraints such as clouds, fog and rain, optical remote sensing has significant limitations, requiring synthetic aperture radar (SAR) remote sensing imaging.
DAMO Academy migrated the optical remote sensing AI algorithm to SAR image analysis by using migration learning technology. Within one week, the algorithm was developed and trained to identify water bodies for critical flood control applications.
Since July, DAMO Academy's remote sensing AI has assisted authorities in completing 262 riverine house identification tasks and 149 water body identification tasks.