One of the things that I find really interesting about this is that it's most likely the best cost/result ratio they could find for sorting tomatoes. The idea that optical recognition, free-falling tomatoes and a series of paddles is the cheapest approach is very impressive.