[toomuchtodo]

Sort of. Machine vision is helpful, but the more trash in the recycling stream, the more dexterity your processing automation requires to handle non-compliant material, as well as the costs of trashing that non-compliant material.

More emphasis needs to be placed on reducing the waste stream as a whole (trash and recyclables), and then leaning on a recycling lifecycle (with the remaining material disposed of with plasma gasification).

[Animats]

Already working in the most advanced recycling plants. Stadler's latest plant has zero people hand-sorting.[1] Here's NYC's big plant in Brooklyn.[2] Sill some hand-sorting. A robotic one from Japan.[3]

It may not be profitable, but it's quite possible to separate just about everything automatically now.

[1] http://www.recyclingtoday.com/article/stadler-installs-fully... [2] https://www.youtube.com/watch?v=nUrBBBs7yzQ [3] https://www.youtube.com/watch?v=HxkklR3BNFc

[toomuchtodo]

Awesome links, thanks for sharing! I won’t call the technical problem “solved”, but the larger issue is the politics that people think recyclables have suffient value to self-sustain recycling operations, and that isn’t the case. You’re going to have to pay for both your trash and recyclables to be disposed of, you’ll just pay less for the recyclables.

[digikata]

Still it seems like this would be a great automation problem to work on - not your standard shiny tech problem but something that would really make a difference.

[toomuchtodo]

I wouldn’t talk someone out of working on solving the waste stream problem.