The use of Machine Learning and Artificial Intelligence in disaster management seeks to help in mitigation wildfires cases. The decision to use these technologies is a move to reduce the devastating effect of wildfires on wildlife and vegetation cover.
The use of Machine Learning and Artificial Intelligence in disaster management seeks to help in mitigation wildfires cases. The decision to use these technologies is a move to reduce the devastating effect of wildfires on wildlife and vegetation cover.
Recent reports show that wildfires in the U.S. are severely destructive because of the rise in global temperatures and changes in weather patterns. Wildfires are natural occurrences, but they become unpredictable and uncontrollable during hot and dry seasons. With this regard, researchers in Stanford plan to use machine learning and satellite imagery to track and predict at-risk areas.
Tests for the susceptibility of forests and scrublands are still conducted manually by sampling branches and foliage to determine their water content. This method is accurate and reliable but becomes labor-intensive and unsuitable for scaling.
All hope is not lost because sources of data are becoming more and more available than before. Sentinel and Landsat satellites are European space agencies that aim to create an accumulative store for imagery of the Earth’s surface. These satellite images are to be keenly analyzed to provide a secondary source of information for wildfire risk assessment. This move intends to reduce splinters and other injuries.
Earlier attempts to implement Earth observation from orbital imagery majorly depended on visual measurement, which is site-specific. The method of analysis differed based on the location. The team at Stanford plans to leverage Sentinel satellites’ synthetic aperture radar that can penetrate forest canopy and take images of the surface.
During a press release, Alexandra Konings said that their latest satellites use longer wavelengths, which help observations to be sensitive to water in the forest canopy. He explained that this aims to be a direct representation of the fuel moisture content. Alexandra Konings is a senior author of the paper, Stanford Ecoydrologist.
The team at Stanford plans to use imagery captured regularly since 2016 and manual measurements made by the U.S. Forest Service, in a machine learning model. This plan allows the model to understand the relationship between specific features of the imagery and the ground-truth measurements.
The team intends to test the resulting Artificial Intelligence program to predict old data for which the answers exist. They plan to give the model up-to-date data to make predictions about future wildfire seasons. With the successful launch of the Artificial Intelligence ‘agent’, authorities can make informed decisions about safety warnings and susceptibility of areas.
Sources:
Apsters News
Provided by the IKCEST Disaster Risk Reduction Knowledge Service System
Comment list ( 0 )
