The United States has created an artificial intelligence system that is being used in rescue efforts in regions of Turkey affected by an earthquake and the aftershocks that followed. This AI system, called xView2, is an open source project that uses a machine learning algorithm on satellite imagery to categorize damage in the hours immediately following a disaster. The US Department of Defense is using the system to provide a faster response to the ongoing disasters in Turkey and Syria. xView2 was developed by the Pentagon’s Defense Innovation Unit and Carnegie Mellon University’s Software Engineering Institute.
The AI system uses the semantic segmentation technique, which examines each individual pixel in a satellite image and its relationship to surrounding pixels, to analyze the state of anything on the ground. After analyzing the image, the system highlights the damage in red, a process that previously could take weeks but is now done in a matter of hours. The AI system has been used in other natural disasters, such as the California wildfires and Nepal floods, to identify damage on the ground and facilitate recovery efforts.
Although xView2 is very efficient at categorizing damage in satellite imagery, its creators point out that there are still problems to be solved, such as its reliance on satellite images taken during the day. This means that the system cannot provide fast data for disasters that occur during the night or in the first hours after a disaster. Despite these challenges, xView2 has proven to be a valuable tool in rescue and recovery efforts following natural disasters.