• Sat. Sep 30th, 2023

all-crypto

news

AI taught to evaluate damage from hurricanes

ByAderinser

Oct 31, 2022

AI taught to evaluate damage from hurricanes

https://gagarin.news/news/australian-pension-funds-expect-clear-regulation-of-cryptocurrencies/

Scientists of Ohio University have developed a machine learning algorithm that predicts the degree of damage to buildings after a hurricane. TechXplore writes about this.

The developers used satellite pictures created before and after a hurricane. These images contained 22,686 buildings.

Then, with the help of a sparkle neural network, the researchers recreated the contours of the buildings in satellite images to the hurricane and classified the amount of damage after the elements. Four categories were used for the model:

  • intact;
  • minor damage;
  • serious damage;
  • Destroyed.

Researchers tested their new model on the Hurricane “Michael”, which fell on the eastern coast of the USA in the fall of 2018. It turned out that the algorithm correctly estimated the damage of 86.3% in one of the regions of Florida. This is 11% better than that of the model created using the support vectors (SVM) method.

“SVM tried to distinguish insignificant and serious damage, which could be a serious problem for groups of response to a hurricane,” said the co -author of the study by Dashen Liu Lyu.

The scientist believes that their algorithm will help rescuers in real time evaluate the damage caused by the element.

“In real hurricane situations, the model can be used to assess the degree of damage to the building to send emergency services there for a priority verification,” Likh said.

Recall that in May 2021, scientists developed a tool with artificial intelligence to predict a city, tornado and strong winds during a storm.

In September, DeepMind introduced a DGMR deep learning tool to predict the rain 90 minutes ahead.

Subscribe to FORKLOG news at Telegram: Forklog AI – all news from the world of AI!

Leave a Reply

Your email address will not be published. Required fields are marked *