An AI-based Automated System for Classifying and Identifying Electronic Waste Using Deep Learning Methods

The increase in electronic and electrical trash (e-waste) is a serious operational and environmental issue. The existing models face challenges to identify different e-waste components from real-time image. It generates the need of Artificial Intelligence (AI)-based automated image recognition system to categorize the e-waste materials. The primary goal of the study is to create a deep learning architecture that can recognize various types of e-waste from image data. Several Convolutional Neural Network (CNN) models were assessed and compared in terms of classification accuracy and loss percentage. Explainable AI was used to identify the major contributory section of the component. Inception V3 and MobileNet were used on the image dataset and the results show that Inception V3 outperforms MobileNet and baseline CNN models with the maximum accuracy of 98.33%. As the system accurately categorizes objects, the recycling processes become easier for practical implementation. This study is innovative as it is an automated system for sorting e-waste components. The system improves the overall processing efficiency and facilitates circular economy implementation. This eco-friendly solution helps the end users and the recycling companies to handle the ewaste in a sustainable approach.