Garbage classification system based on artificial intelligence and Internet of Things

Abstract

The existing garbage classification system has some problems such as poor classification effect, high classification cost and difficult management. In order to solve these problems, this paper proposes a garbage classification system that can manage multiple garbage cans simultaneously and classify them accurately in real time. The system deploys the improved MobileNet V2 network model on the low-cost K210 module, and adopts the method of first object recognition and then garbage classification to classify the garbage images read by the camera. At the same time, the system collects environmental information such as temperature and humidity, overflow situation, and air quality of the garbage cans through sensors, then summarizes the classification information and environmental information at the terminal, and finally uploads the aggregated information to the cloud through the NB module. After physical tests, the system can classify garbage with an accuracy rate of more than 90%, run at a maximum speed of 12 frames per second, and can monitor environmental information in real time, which has certain practical value.