A Novel Hybrid Method for Vision‐Based Target‐Free Dynamic Displacement Measurement of Civil Structures Under Seismic Excitation

ABSTRACT Dynamic displacement measurement is critical for assessing the structural integrity of civil structures in the aftermath of seismic events. However, traditional contact sensors and current non‐contact vision‐based methods face persistent challenges in accuracy, deployment flexibility, and robustness under complex field conditions. To address these limitations, this paper proposes a novel hybrid method for vision‐based target‐free dynamic displacement measurement of civil structures under seismic excitation. Departing from single‐algorithm approaches, the proposed method synergistically integrates template matching (TM) and optical flow (OF) through a dynamic interaction and fusion mechanism. The TM channel delivers robust pixel‐level displacement estimation while the OF channel provides sensitive sub‐pixel velocity. These channels interact dynamically, where TM‐derived displacements dynamically optimize the OF window size in response to structural motion to balance measurement sensitivity and robustness. The system further employs Kalman filtering technique to fuse channel outputs, leveraging OF for a priori motion prediction and TM for a posteriori error correction, thereby mitigating individual limitations such as TM's quantization errors and OF's cumulative drift. The proposed method is experimentally validated on an outdoor cold‐formed steel (CFS) wall system and further applied to the engineering scenario of dynamic behavior analysis of reinforced concrete (RC) building. The results have shown that the proposed method provides more accurate, robust and efficient dynamic displacement measurements and overcomes the limitations of single‐channel measurements, thus making it promising for structural health monitoring (SHM) in civil engineering.