Technology

Australian university develops drone that spots Covid-19 in crowds

31 March 2020

The University of South Australia (UniSA), in partnership with Canadian tech firm Draganfly, is developing a “pandemic drone” designed to detect people with Covid-19 and other respiratory diseases from above.

The drone will use sensors that can monitor people’s temperature, pulse and breathing, and can detect sneezing and coughing in crowds. It can carry out its mass diagnosis from a distance of up to 10m and can be used indoors or out. Fixed cameras can do the same at a distance of 50m.

The UniSA team is led by Javaan Chahl, professor of sensor systems. He was part of a research group that demonstrated in 2017 how a drone could use image-processing algorithms to detect heart rates.

“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,” he said.

The technology was originally developed to help medical teams in war zones and natural disasters, and for monitoring premature babies in incubators.

“Now, shockingly, we see a need for its use immediately, to help save lives in the biggest health catastrophe the world has experienced in the past 100 years,” said Chahl.

Cameron Chell, chief executive of Draganfly, said in an interview last month that facial recognition software could be added to the data package to identify potentially infected individuals. However, he said the purpose of the flights would be to identify the pervasiveness of an infection within a population, rather than picking out specific individuals.

Suitable hardware platforms include Draganfly’s Commander quadrotor. Indoor flights could be carried out with drones using ducted fans, in which the rotating parts are shielded, to make them safer in the event of a malfunction.

Chell said: “These are a bit bigger than your typical hobbyist drone, but they can still be small enough to fit inside a backpack.”

Image: Drone-diagnosis may be faster and cheaper than using fixed cameras (Nevit Dilmen/CC BY-SA 3.0)