YOLOv8-Based Real-Time Detection of Recyclable Metal, Plastic and E-Waste for Environmental Waste Segregation

Efficient segregation of recyclable materials is critical for improving municipal solid waste management and supporting circular economy initiatives.This paper presents a real-time waste-sorting prototype based on the YOLOv8 object detection model to identify recyclable metal, plastic and e-waste items from live video using a standard webcam, together with a web-based interface to visualize and log detections.A custom dataset of labelled waste images, collected under realistic lighting and background conditions, was used to train and evaluate the model.Experimental results show high detection performance for metal and plastic, and moderate performance for e-waste, while maintaining real-time operation suitable for deployment in material recovery facilities, collection centres and smart bins.The proposed system demonstrates how AI-enabled waste-segregation technology can support more efficient recycling operations and generate management-relevant data for environmental decision making.