Utilizing Agentic AI for Rebar Tracking on Construction Sites to Enhance Construction Project Management

Abstract

Accurate construction material management is essential for enhancing project efficiency and minimizing unnecessary waste. This study proposes a rebar inspection system that integrates Agentic Artificial Intelligence with Computer Vision technology to enable automated tracking and control of construction materials. The system follows a cyclical workflow consisting of planning, decision-making, and performance evaluation, using real-site imagery to inspect rebar across key stages—ranging from transportation to placement before concrete pouring. The YOLOv8 model is utilized for detecting both rebar cross-sections and full-length geometry, with results compared against a 3D structural model to verify alignment with actual material requirements. The acquired data is further processed to optimize rebar cutting patterns based on standard commercial lengths, aiming to reduce waste and improve material utilization. The implementation of Agentic AI in this context demonstrates strong potential for advancing construction material management systems, enhancing both operational efficiency and readiness for deployment in real-world construction environments.