Page 150 - AI for Good-Innovate for Impact Final Report 2024

AI for Good-Innovate for Impact

• EF-UC33-REQ-003: It is critical that the project employs community empowering
technologies for early detection of fires, as early detection is key to extinguishing fires.
This includes the use of computer vision (cameras based on python, identify smoke, 20
km range) and satellites.
• EF-UC33-REQ-004: The project must utilize on-premises computers and outdoor
infrastructure to manage wildfire risk effectively. It should leverage NASA’s geographic
positioning data of communities, power lines, and other human infrastructures, such
as roads and villages. Additionally, the project must consider the limitations of satellite
detection, including latency between detection and action, the diverse capabilities of
satellites, and the need for coordination of satellite orbits.
• EF-UC33-REQ-005: It is critical that the project uses anomaly detection for smoke and
light, and leverages streamed videos, detect smoke and light, Satellite images (used only
for post-fires), and Sensors from satellite which can detect heat (diverse generation of
sensor) for model training and fine tuning.

33�4� Sequence diagram

33�5� References

[1] Measuring Gross Carbon Dioxide Emissions from Forest Fires in the Abrace o Pantanal
Project Area

134