Poachers typically hunt at night, which is why drone-mounted infrared cameras are being used to spot them. The problem is, since both the poachers and the animals emit heat, it can be difficult to tell which is which in the videos. Scientists from the University of Southern California are making the job easier, using artificial intelligence.

Typically, wildlife officers have to spend the night with laptop computers at base stations, monitoring infrared video being streamed from drones. When a heat-emitting blob appears on the screen, they then have to figure out if it's a human or an animal, which isn't always easy.

That's where SPOT comes in. An acronym for Systematic POacher deTector, the algorithm was developed by a team led by computer science PhD student Elizabeth Bondi.

The researchers started by digitally labelling 180,000 humans and animals in infrared videos. Using that dataset and a modified version of an existing deep learning algorithm known as Faster RCNN, they then taught a computer how to distinguish between the two types of images. While it could do so accurately, it was taking 10 seconds to process each image, which is too long when the footage is being shot by a moving drone.

With that in mind, the team altered the algorithm so that it could work with the Microsoft Azure cloud computing platform. Utilizing Azure's faster processing time, SPOT is currently able to tell poachers from animals in about three-tenths of a second.

Plans now call for a large-scale deployment of the technology across Botswana. Another AI-based poacher-spotting system was recently developed by tech firm Neurala.