UAV-based system uses sensors and AI to detect symptoms of respiratory illness in crowds.

The computer vision and sensor system will be able to detect sneezing and coughing actions of individuals in crowds, as well as detecting heart rate and respiratory rates and temperature.

The researchers suggest it could be deployed in areas of congregation that present high risk of infection, such as offices, cruise ships, airports, and sporting matches.

Dr. Javaan Chahl, Defence Chair of Sensor Systems at UniSA, said the system could be a screening tool for Covid-19.

“It might not detect all cases, but it could be a reliable tool to detect the presence of the disease in a place or in a group of people.”

“There’s a lot of engineering going on right now but the aspiration is to have this in some sort of initial capability within six months,” he said.

“It’s one thing to have it work in a science experiment type scenario but getting it to run in the field on a real piece of hardware is quite a challenge.”

In 2017, Professor Chahl’s team demonstrated that a human’s heart rate could be extracted from a video with image-processing algorithms. The technology has developed since to measure breathing rate with high accuracy, and interpret actions such as sneezing and coughing.

The new project is being developed in partnership with Canadian firm Draganfly.

Stay up to date by getting stories like this delivered to your mailbox.
Sign up to receive our free weekly Spatial Source newsletter.