Researchers are using artificial intelligence to turn poachers into prey.

The University of Southern California Center for Artificial Intelligence in Society applied deep learning to spot hunters in near real time.

What began with impoverished peasants looking for food has grown into a global dilemma, one that poses a serious threat to the survival of many plant and animal populations.

USC has long applied AI to protect wildlife; analysts initially used artificial intelligence and game theory to anticipate poachers’ haunts. Now they have the technology to stop them in their tracks (so to speak).

Tools like infrared cameras are great for monitoring poachers during nighttime hunts. Except, of course, that the equipment picks up human and animal heat maps, making it difficult to focus on the right mark.

So, a team of USC computer scientists employed their proprietary labeling tool to tag 180,000 people and critters in infrared videos. And with the help of a modified version of the Faster RCNN deep learning algorithm, the team was able to teach a computer to automatically distinguish infrared images of humans from those of animals.

The challenge, then, was to deploy the algorithm to spot poachers, recorded live by drones hovering above.

Put to the test at national parks in Zimbabwe and Malawi, the algorithm took 10 seconds to process each image—too long when an elephant, tiger, rhinoceros, or gorilla’s life is on the line.

With a few tweaks (including a move to Microsoft’s Azure cloud computing platform) “SPOT”—the Systematic POacher deTector—now works to detect poachers and animals in about three-tenths of a second.

“SPOT will ease the burden on those using drones for anti-poaching by automatically detecting people and animals in infrared imagery, and by providing detections in near real time,” lead author Elizabeth Bondi, a Ph.D. candidate in computer science at USC, said in a statement.

This sounds a lot more high tech than some of the other tactics, like robotic decoy animals, used to lure poachers into a trap.


Let us know what you like about Geek by taking our survey.




Please enter your comment!
Please enter your name here