TY - JOUR
KW - Construction progress monitoring
KW - Percentage of completion
KW - Construction activity identification
KW - Depth estimation
KW - Bridge piers
AU - Yuyang Wang
AU - Elton Chen
AU - Mirosław Skibniewski
AU - Zhe Wen
AU - Zehao Jiang
AB - Effective construction progress monitoring (CPM) is essential for successful engineering projects. However, given the large number of components, wide spans, extended construction cycles, and section-by-section pouring involved in pier group construction, existing approaches struggle to quantify progress effectively. To overcome this, this paper proposes an unmanned-aerial-vehicle-based approach to evaluate both percentage of completion (POC) and construction activity at reduced cost. A depth estimation model tailored for large-scale construction scenes (LSC-Depth) is developed to address inaccuracies in weak-texture regions and those with absent structural boundaries. Construction progress is quantified by computing bridge pier height through geometric mapping. Simultaneously, construction activities are inferred from the spatial and logical relationships among site elements. The proposed approach is validated through application to an actual pier construction project. It achieves accuracies of 0.95 for POC and 0.85 for activity-level CPM. Future research shall focus on improving POC accuracy and refining activity-level granularity to enhance managerial efficiency.
BT - Automation in Construction
DO - 10.1016/j.autcon.2026.107040
N2 - Effective construction progress monitoring (CPM) is essential for successful engineering projects. However, given the large number of components, wide spans, extended construction cycles, and section-by-section pouring involved in pier group construction, existing approaches struggle to quantify progress effectively. To overcome this, this paper proposes an unmanned-aerial-vehicle-based approach to evaluate both percentage of completion (POC) and construction activity at reduced cost. A depth estimation model tailored for large-scale construction scenes (LSC-Depth) is developed to address inaccuracies in weak-texture regions and those with absent structural boundaries. Construction progress is quantified by computing bridge pier height through geometric mapping. Simultaneously, construction activities are inferred from the spatial and logical relationships among site elements. The proposed approach is validated through application to an actual pier construction project. It achieves accuracies of 0.95 for POC and 0.85 for activity-level CPM. Future research shall focus on improving POC accuracy and refining activity-level granularity to enhance managerial efficiency.
PY - 2026
EP - 107040
T2 - Automation in Construction
TI - Automated monitoring of percentage of completion and activity-level progress in bridge pier construction using unmanned aerial vehicles
UR - https://www.sciencedirect.com/science/article/pii/S0926580526002815
VL - 188
SN - 0926-5805
ER -