Researchers at Johannes Kepler University in Austria have created an autonomous drone that can help with search and rescue missions for lost people in hard-to-reach places, such as in very dense forests or mountains where the field of vision is extremely restricted.
The new technology uses special algorithms implanted in a camera system that facilitates observation through the treetops. In addition, the device is able to map and highlight people, stationary or moving, who are under large areas of dense forest with reduced visualization.
Unlike first responders in planes and helicopters, who have trouble seeing under tree cover, or thermal sensors, unable to take accurate readings under vegetation, drones have become a viable option for finding lost or injured victims.
Even so, they have certain limitations, as they are remotely controlled by pilots who face the same problems in viewing the area just below the treetops. With the new technology, it is easier to identify the exact place where one or several people are.
Optical Cutting Algorithm
This system uses the processing power of a computer to blur objects that can obstruct vision, such as treetops. Thermal imaging helps highlight the heat emitted by the body while a machine learning application analyzes whether the signals belong to humans, animals or other heat sources.
The device works on a conventional autonomous drone that uses a spatial positioning system to determine which area to survey. When the sensors find something, the drone approaches the location for a more accurate view and sends a signal to the nearest rescue team.
In the first 17 tests carried out away from the laboratory, the drone equipped with the optical cutting algorithm managed to locate 38 of the 42 people hiding under the treetops. In addition, the system was able to accurately identify in real time, two fundamental qualities in search and rescue missions.
help comes from heaven
This novelty somehow resembles a project developed in Germany, where researchers created a system that uses drones to identify the sounds of buried people screaming for help or trapped in debris in disaster areas. Unmanned aerial vehicles are equipped with special microphones that pick up different noises, using different types of filters connected to a machine learning central.
In open-field tests, the interval between the 60-microphone system picking up the sound of a scream and accurately identifying its location to the rescuers was just a few seconds. This agility was also observed in situations with excessive external noise and with several people screaming at the same time.
The post Drones use algorithms to locate people lost in forests appeared first on CmaTrends.