Despite the importance of construction and demolition waste (CDW) management, conventional weight-based measurement approach is limited because its complicated heterogeneous characteristics make accurate weight estimation for different types of CDW difficult. This study develops an AI-assisted quantitative analysis system for instant estimation of CDW treatment costs at construction sites. The system, trained on 3,044 CDW images with the YOLOv8s instance segmentation model, achieved high detection accuracy (mAP@0.5 = 0.98). Yet, performance declined for reflective or irregular waste, highlighting the need for expanded datasets and field validation. Additionally, a dual-structured estimation framework was developed to reflect the needs of both contractors and intermediate processors. This framework was then standardized into a database and implemented as a web-based application to enable real-time information delivery. The proposed system demonstrated high accuracy—particularly for concrete waste—and successfully provided immediate cost estimation capabilities in field conditions. This study presents an automated and transparent solution that minimizes unnecessary transport, redundant processing, and material loss through AI-driven volume estimation. The system enhances the efficiency and transparency of CDW management and contributes to sustainable resource circulation and net-zero construction strategies.
Chang et al. (2025) studied this question.