Machine learning fulfills its civic duty by looking at roadside garbage

If there’s one thing that never seems to be a supply chain problem, it’s litter. It’s everywhere, easily identifiable and – you think – pick it up Sadly, most of us seem to consider garbage as someone else’s problem, but with something like this machine vision litter mapper, you can at least be part of the solution.

For civic minded people [Nathaniel Felleke], The litter problem was getting too much in his native San Diego. He argued that a map of where the trash was located could help municipal crews clean up, so he suggested creating a system to automatically search for trash. Using Edge Impulse and a collection of roadside images captured from various sources, he created a model for garbage detection. To find trash, a webcam with car window mounts captures images while driving and runs a Raspberry Pi 4 model and searches for trash. When garbage is found on the side of the road, Pi sends the GPS location of the garbage to a cloud database via its cellular modem using a blues wireless notecard.

Wandering the streets of San Diego, [Nathaniel]Its system creates a database of garbage hotspots. From there, it’s easy to drag data to create a heatmap of where the garbage is and overlay it on Google Maps. The video below shows his system in action.

Yes, driving around a private vehicle specifically to spot garbage is only adding more waste to the mix, but you would imagine putting something in municipal vehicles that is already driving around cities. Anyway, at least we didn’t go down without explaining ourselves first. We’ve seen them used before, but [Nathaniel]Its project gives us a way to move forward with some ideas that we have been wandering around for a while.

Leave a Reply

Your email address will not be published.