Design and Implementation of an Autonomous Waste Sorting Machine Using Machine Learning Technique

Abstract

In a world where the population is inevitably
increasing, waste produced is progressively increasing as well. In
this project, an autonomous waste sorting machine was made
which could detect multiple classes of waste materials, and then
separate them accordingly. The waste products were taken as
input in a funnel-shaped structure and dropped one by one to a
conveyor belt where they would be detected by machine learning
technique using Faster-RCNN, and then a servo motor would
separate them according to the detection result. In rare cases,
there are some misdetections of the waste materials, but the
reliability of the detection was very high. Our work can facilitate
human efforts to separate waste products and can make the
waste sorting system completely automatic.