Nvidia engineer uses AI to stop cat from pooping on his lawn

Technology giant nVidia has revealed how one of its engineers has used machine learning at his home to keep cats off his prized lawn.

Robert Bond, who has previously used the firm’s Jetson TX1 platform to build a laser to take out the ants that appeared on his kitchen floor, used the same machine learning technology to turn his sprinklers into a smart identification system that could spot cats that appeared on or near the lawn before triggering the sprinklers to shoo the feline visitors away.

“My wife is a gardener and she likes her garden to be tidy and clean,” Robert said. “It wasn’t actually that much work. The new Jetson TX1 is really good at running these neural nets.”

His system, as he explained in an nVidia blog post, works by detecting motion. Once it is detected, a camera aimed at Robert’s front garden captures an image. Intelligent software running on Jetson then scans the image for any signs of shapes it can identify as cats.

To train it, Robert fed hundreds of images of cats into the software. Once identified, it triggers a chip board he’d soldered to his irrigation control box and turns the system on, scaring cats away.

Robert also revealed that he’s working on a system to help pinpoint the location of cats detected in images, with the possible aim of deploying a remote control car to help get rid of the invaders.

Source: SNAP.PA