This post is also available in: heעברית (Hebrew)

A Team of experts from Malaysia suggests a practical solution for search and rescue operations that involves a real-time human detection system using a fixed-wing Unmanned Aerial Vehicle (UAV).

Cheok Jun Hong and Vimal Rau Aparow of the University of Nottingham Malaysia created UAV technology with readily available small-scale tools (like the small single-board Raspberry Pi computer), allowing both to better manage system functions than with conventional technology and stream aerial imagery from an attached camera.

According to Techxplore, this approach is especially useful due to its ability to delegate computationally intensive human detection tasks to a server at the edge, which is enabled by 4G cellular network technology. The team explains that the server uses a deep neural network that can precisely identify people from the images gathered by the UAV’s camera and transmit results to ground control. After receiving the identification, a rescue team can then be sent to the exact spot where a rescue is needed.

The system combines deep learning algorithms with mobile-edge computing, thus representing a shift away from conventional search and rescue approaches that could speed up the whole process during crises.

The team explained that their convolutional neural network with YOLOv3 architecture can identify people in the images from the UAV camera almost 80% of the time. The researchers can further optimize the approach by using the TensorRT toolkit and speed up the inference without loss of accuracy. Nevertheless, while the system has a better range than one that is radio-enabled, it does rely on the stability and existence of the 4G network across the search and rescue area.

Finally, while the researchers initially designed the system for human search and rescue scenarios, it could also be adapted to other applications like public safety and crime prevention, patrolling a site vulnerable to criminal activity or even used in tracking criminals.