Researchers at QUT, the Queensland University of Technology in Australia and the Brisbane City Council are using drones with heat-seeking sensors to relate the Brisbane’s koala population despite the concealing foliage.
The high tech effort by the Queensland University of Technology specifies, that the experts will be able to receive more accurate estimates of the koala population, which will support the team to produce effective conservation decisions.
The research team believes that this method could be more efficient than identifying and mapping koalas and other endangered species rather than hiking via bushland.
QUT’s Grant Hamilton and colleagues have designed a method of using computer algorithms to understand the data from the heat-seeking sensors. Hamilton says, “We feed that thermal imagery into an artificial intelligence algorithm and we use those algorithms to automatically determine the kind of organisms present in the environment.”
The idea is to have two drones which can fly for about 30 minutes. They contain thermal imaging cameras associated with them, as well as a high-resolution camera.
He also says the system is returning information with 100% accuracy in certain environments. “The drones will be flown in an accurate way to permit us to cover the area and to maximize the probability of detection. Then they will carry the information back to us and we will be working on AI algorithms to find all the koalas”.
Hamilton mentions, “At first light, the drones will be flown around dawn, which is the best time to wait for koalas since it highlights the differences in temperatures. It’s hard to analyze the heat of something that’s sitting on the surface of the identical temperature. For instance, you can see a human on a cold day standing on ice, but if they were in the desert on a hot day, it would be difficult to see them. So by flying the drones in cool weather, around dawn, means it will be painless to detect the koalas.”.
This estimation is the first of its kind in developing a robust methodology to use drones to identify the numbers, accounting for errors in detection.
Watch a video of heat seeking Drones in action: