AI-based systems are designed to detect poachers

Hunters often hunt at night, which is why infrared cameras mounted on unmanned aircraft are used to detect them. The problem is, because both poachers and animals emit heat, it is difficult to accurately identify. Scientists from the University of Southern California are making this easier by using artificial intelligence on a system called SPOT.

Normally, wild animal guards must sleep at night with laptops at guard stations, watching infrared videos transmitted directly from unmanned aircraft. When the heat spots appear on the screen, they have to determine whether it is a human or a wild animal, but this is not easy.

That's why SOPT was created. SPOT is an acronym for Systematic POacher deTector, an algorithm system developed by the research team conducted by Elizabeth Bondi, a computer science doctor.

The researchers began by integrating 180,000 human and animal digital labels in infrared videos. Using that data set and creating a modified version of the existing deep learning algorithm, called Faster RCNN, then they taught computers how to distinguish between two types of thieves and animations. object. Although the system algorithm can correctly identify, it takes 10 seconds to process each image, but this time is too long.

With that in mind, the team changed the algorithm so that it could work with Microsoft Azure's cloud computing platform. Using Azure's faster processing time, SPOT is now able to identify poachers exactly in about three-tenths of a second.

Current plans require extensive deployment of technology across Botswana. Another system based on artificial intelligence was developed by Neurala.

See more:

1. AI can take away the job but will bring a better life
2. This is how technology shapes education in the future
3. Microsoft Research is creating an AI capable of writing code in seconds