In times of crisis, such as fire or tsunami, rescue robots aid human rescuers to access hazardous places. Recently, thanks to advances in AI techniques, rescue robots have gained the ability to rescue people autonomously, simply with just sensors and cameras. 

Rescue robots can autonomously create a map of the area and identify survivors. They obtain a series of 2D images from the camera, which are interpreted using computer vision techniques. Simply, robots identify key features such as corners and edges from different angles, during which 2D points are converted into 3D points. This is repeated for multiple key features, ultimately constructing a 3D map of the environment. This way, robots can avoid obstacles and navigate the area efficiently. Most of all, such 3D maps are great sources for human rescuers to learn the interior structure, particularly if the building has collapsed.

https://robohub.org/wp-content/uploads/2019/08/Image2.png

Moreover, images are also used to identify humans and take action. From the series of images, robots can locate and distinguish humans from obstacles. Robots can even analyze certain poses of the survivors to take appropriate action. For instance, if a person is waving their hands, the robots can interpret the images using Neural Network to identify the person’s action in the image. Then, the robot can notice the person is calling for help and set the priority to rescuing the person. 

In essence, rescue robots are indispensable tools in modern emergency response. Ultimately, rescue robots can provide swift and efficient support in times of crisis
.
 

인수호 강남포스트 학생기자  webmaster@ignnews.kr

<저작권자 © 강남포스트, 무단 전재 및 재배포 금지>