A novel integration of hyper-spectral imaging and neural networks to process waste electrical and electronic plastics

Abstract

In this study, a technique which combines hyper-spectral imaging technology and a neural networks-based algorithm has been introduced for identification and separation of different types of e-waste plastics (e-plastics). Although recent technological developments in computing power allows for the handling of big data in a relatively reasonable time, a manageable number of neurons must be utilized in order to realize real-time sorting applications for plastic recycling. A successful result to identify three different common types of e-plastics with a very high rate of accuracy has been presented. The result has been achieved using a special designed Artificial Neural Networks (ANN) algorithm and hyper-spectral signature of those plastics. The promising result will pave a road to address the shortcomings of current e-plastic sorting technologies in terms of efficiency and reliability.