Convolutional neural networks application in plastic waste recognition and sorting

In this paper authors describe their results in IoT and Convolutional Neural Networks application in reverse vending machine project. In Europe and United States a reverse vending machine is a device that accepts used (empty) beverage containers and returns money to the user. In Russia we can provide the concept of automatic processing of plastic and metal waste combined with another mechanism for economic motivation of end users - system of discounts and bonuses in large trading networks. These machines have to be as cheap as possible but with rich functionality and security. Traditionally the basics of this machine is assembled with IoT controllers and tiny single-board computers which have heavy memory and computational restrictions. Nevertheless these controllers are able to recognize the waste types using cameras and provide sorting and some kind of preprocessing. We show our CNN implementation on IoT for reverse vending machine.