Application of Thermal Imaging for Search and Rescue Missions Using Unmanned Aerial Systems
Disciplines
Other Computer Engineering | Robotics
Abstract (300 words maximum)
The use of Unmanned Aerial Systems in recent years has been a constantly evolving technology. It has become very common for various groups to use UASs to complete tasks in a much more efficient way than was possible in years past. Flying machines can be utilized in ways that traditional human efforts can’t and can be much more cost effective. The method that will be highlighted in this paper is the use of UASs in search and rescue operations, combined with thermal and infrared imaging. In an emergency, such as when hikers get lost in the forest, time is of the essence and a quick response time can be the difference between life and death. However, these situations can be difficult to maneuver using traditional means. Traditional rescue missions heavily relied on manpower and ground equipment to complete the task, not to mention being extremely costly. Thermal imaging has become a crucial method; however, there can be problems using this method as well. With a standard visible light camera and even a thermal imaging camera, there can be problems when the conditions become difficult to see through The purpose of this paper is to determine how accurately our UAS combined with Ultralytics YOLO (You Only Look Once) and OpenCV, can detect human beings and determine the forehead temperature of the person in the image. We took multiple, non-contact radiometric images and compared them using Flyr to compute a per-pixel temperature map (°C). Our comparison was between two pathways of YOLOv10 and YOLOv12 to create segmentation for the human head vs Pose estimation for the human forehead. Our results provided qualitative visualizations showing a person's silhouettes, head outlines, and temperature readings. Outputs are saved to the respective approach folders. When radiometric data is present, the overlay reports temperatures in °C.
Use of AI Disclaimer
no
Academic department under which the project should be listed
SPCEET – Robotics and Mechatronics Engineering
Primary Investigator (PI) Name
Adeel Khalid
Application of Thermal Imaging for Search and Rescue Missions Using Unmanned Aerial Systems
The use of Unmanned Aerial Systems in recent years has been a constantly evolving technology. It has become very common for various groups to use UASs to complete tasks in a much more efficient way than was possible in years past. Flying machines can be utilized in ways that traditional human efforts can’t and can be much more cost effective. The method that will be highlighted in this paper is the use of UASs in search and rescue operations, combined with thermal and infrared imaging. In an emergency, such as when hikers get lost in the forest, time is of the essence and a quick response time can be the difference between life and death. However, these situations can be difficult to maneuver using traditional means. Traditional rescue missions heavily relied on manpower and ground equipment to complete the task, not to mention being extremely costly. Thermal imaging has become a crucial method; however, there can be problems using this method as well. With a standard visible light camera and even a thermal imaging camera, there can be problems when the conditions become difficult to see through The purpose of this paper is to determine how accurately our UAS combined with Ultralytics YOLO (You Only Look Once) and OpenCV, can detect human beings and determine the forehead temperature of the person in the image. We took multiple, non-contact radiometric images and compared them using Flyr to compute a per-pixel temperature map (°C). Our comparison was between two pathways of YOLOv10 and YOLOv12 to create segmentation for the human head vs Pose estimation for the human forehead. Our results provided qualitative visualizations showing a person's silhouettes, head outlines, and temperature readings. Outputs are saved to the respective approach folders. When radiometric data is present, the overlay reports temperatures in °C.