Paper Title
Towards Sustainable Urban Development: Comparative Study of YOLO Model Weights on Construction Debris Detection

Abstract
Efficiently managing construction and demolition waste is crucial for sustainable urban progress and aids in the realization of numerous Sustainable Development Goals (SDGs). This paper suggests a deep learning-oriented method for identifying and categorizing construction debris utilizing the Construction and Demolition Waste Dataset (CODD) and various YOLO (You Only Look Once) model structures. By comparing and assessing the performance of different YOLO models, in- cluding lightweight and high-capacity versions, their effectiveness is evaluated based on accuracy, computational efficiency, and adaptability to diverse environmental situations. The analysis centers on precision, recall, and mean Average Precision (mAP) as essential performance indicators. The study underscores the flex- ibility of YOLO models for automated debris detection activities, highlighting their potential for incorporation into scalable waste management frameworks. This research illustrates the impactful role of deep learning in tackling practical challenges in waste management, making a significant contribution to environmental sustainability and resource optimization. Keywords - Construction Debris Detection, Sustainable Development Goals (SDGs), Deep Learning, Object Detection, Waste Management