An Application of a Robot Arm for Garbage Classification Using Image Processing and Artificial Intelligence

Garbage classification is an important problem in big cities. By using a combination of free tools available from the open community, a platform can be created in laboratories for simplify this complicated systems into a simple model for students. However, this system still maintains rich features that help students in researching. In this paper, we present a platform that satisfies these requirements. Besides creating a mechanical robot arm Dobot M1, forward and reverse kinematic calculations are examined for this model. Thence, through image processing combination, we apply a Yolo neuron network (NN) structure. By using this network, based on a set of available data, we enrich the date by our new data. This structure is trained to classify kinds of garbage. Our experimental model works successfully through a survey to prove the high application of this research. A supervision platform is also presented for operators to control the process.