It seems like an odd paradox. Astronomers can see stars billions of lightyears away and yet conservationists cannot see animals right here on earth.
Two friends and colleagues at Liverpool John Moores University recognised this and decided to team up to create a new technology for surveying endangered species as well as their poachers. This technology uses drones equipped with infrared cameras and combines them with computer-vision machine-learning techniques employed by astronomers.
The team first created a ‘thermal library’ by testing their drones in a safari, collecting thermal footprints from thousands of animals and using this data to improve algorithms. Then, they tested their technology in England, using it to spot cows as well as fake “poachers” or rather students who were hiding in the bushes around the site. In Australia, they tested for the ideal flight height, developed software to account for the thermal effects of vegetation, and improved algorithms that differentiate animals and people from heated inanimate objects. Finally, the team took their research to South Africa, where they successfully spotted five Riverine rabbits, an extremely endangered species that has only been spotted about 1000 times in the wild.
This technology has the potential to help not just ecologists, but conservationists as well. The current method for counting animals relies on physical sightings and signs that the animals leave behind, which is both time consuming and inaccurate. Drone technology has the power to change this. Additionally, since poachers hunt mainly at night, typical optical cameras cannot detect them, but even under the veil of darkness, poachers cannot hide their thermal footprint.
This month, the team will travel to Malaysia to search for orangutans, then to Mexico to seek spider monkeys, and then out to Brazil to spot dolphins.
Image: Conservation drone in action. Liverpool John Moores University
This story is taken from the 13 April 2018 edition of The Warren Centre’s Prototype newsletter. Sign up for the Prototype here.