Waste Classification and Monitoring Using Yolov8 A Deep Learning Object Detection Technique

With rapid urbanization and population growth, effective waste management has become an important issue worldwide. Increasing waste generation poses serious risks to the environment, economy, and public health, necessitating the adoption of best waste management practices. This research proposes a new method for object detection using YOLOv8, the latest version of the “You Only Look Once” (YOLO) real-time object detection algorithm. YOLOv8 was selected for its ability to quickly and accurately identify different types of waste, including plastic, metal, and organic waste, which is crucial for improving the recycling process and promoting a safer environment. However, our preliminary findings indicate that the accuracy of the classification function (mAP50: 59%) is lower than expected. This issue may stem from factors such as limited diversity in the training data, insufficient data for some specific types of waste, and challenges in identifying common objects within images. By addressing these limitations, we aim to improve the model’s performance, making it more effective for waste management. Our model can identify and classify waste products, such as plastic, metal, paper, and organic waste, using deep learning and neural networks. However, while maintaining overlap accuracy, it appears slightly less competitive. Despite these limitations, the model demonstrates potential to enhance waste management processes by increasing the speed and accuracy of waste identification and classification. Its accessibility offers a cost-effective and scalable tool for real-time waste management. This research will contribute to knowledge creation in sustainable urban waste management, paving the way for improved environmental practices and public health protection.