Just-in-time construction progress management for slope engineering with unmanned aerial vehicle-based automated spatial analysis

Abstract Effective progress monitoring is crucial for optimizing the management of construction projects, particularly in slope engineering, where timely and accurate tracking of progress can significantly enhance efficiency. Traditional methods of monitoring often rely on manual intervention, leading to delays, inaccuracies, and high labor costs. This paper presents a just-in-time approach for automatic construction progress tracking using unmanned aerial vehicle-based spatial analysis, which minimizes human input and maximizes real-time monitoring. The proposed framework integrates an innovative point cloud segmentation technique, a virtual multi-view point cloud dense reconstruction method, and a voxelization process to convert unordered point cloud data into structured, actionable information. We applied this framework to Huangmeishan Tunnel Expansion Project in China, processing large-volume data sets. The automatic segmentation method achieved an average model accuracy of 96.55% compared to manual segmentation. Additionally, the automatic excavation volume calculation showed a 97.50% accuracy. By converting point cloud surface models into solid models and organizing them into structured sequences, the framework facilitates rapid spatial information analysis and provides real-time insights into project progress, including quantity estimation, progress monitoring, and over-excavation/under-excavation analysis. This automated approach ensures timely and efficient monitoring, offering a robust solution for JIT progress tracking in slope engineering.