Displacement measurement of underground prefabricated metro stations by computer vision technology

The increasing applications and structural vulnerability of prefabricated metro stations have resulted in a growing demand for monitoring their displacements during the construction and post-construction phases. This paper proposes a dual-camera computer vision monitoring system consisting of a dual-camera base station, a series of monitoring prisms and two reference prisms. The displacement and rotation of the camera station are considered to reduce measurement errors caused by camera motion. The system has been applied in the displacement monitoring of the 164 m assembly section of an underground station in Shenzhen during the construction and post-construction periods. A movable target equipped with a spiral micrometer is used for field calibration experiments, and the results show that the average measurement deviation is 0.01 mm and the standard deviation is 0.068 mm. During the 5-month monitoring, displacement caused by backfilling, rainfall, and other construction conditions are precisely reflected, which shows a potential correlation between uneven displacement and some minor structural defects. At the post-construction period, settlement variance is relatively small, and the temperature effect causes horizontal convergence to fluctuate.