AI-Driven Multi-Class Waste Classification for Sustainable Recycling Infrastructure in Post-Conflict Zones Using CNN, IoT, and AR

Abstract

Regions affected by conflict leads to increasing amount of waste accumulation. This has a profound impact that goes beyond the geographical boundaries. The restoration process comes with a lot of challenges. For environmental restoration and reconstruction, we need efficient waste recycling and debris management strategies. This research proposes a strategic model for recycling and reuse in post-conflict zones. The AI-powered deep learning model aims to classify mixed household waste by utilizing a dataset containing 15,150 images from 12 household garbage categories. The model's predictions make use of Convolutional neural network (CNN) to train and classify waste with high accuracy. The trained model can segregate waste into recyclable, biodegradable, and non recyclable groups. Our solution aims to foster sustainability and assist in urban recovery by serving as a guide to strategic planning for mobile recovery units and local recycling networks. By integrating the proposed approach with IoT devices and augmented reality (AR) interfaces, this study outlines a smart material recovery system optimized for waste management of construction debris, thereby being a generally applicable solution for post-disaster environments.