Computer vision‐based monitoring of the 3‐D structural deformation of an ancient structure induced by shield tunneling construction

Abstract

Summary Shield tunneling is an efficient method of the tunnel construction for the system of underground transportation, but it will disturb the surrounding soil and affect the security of nearby structures. The monitoring for deformation of the structures influenced by the tunneling construction is crucial for structural safety evaluation, especially for the ancient structures. The computer vision‐based structural deformation monitoring approach is a newly developed method that has advantages such as, long‐distance, noncontact, and so on, which provides a promising tool for deformation monitoring of ancient structures. This study applied a computer vision‐based method for field monitoring of three‐dimensional (3‐D) deformation of an ancient tower under the influence of shield tunneling. The target‐free strategy was adopted to eliminate the need for installation of artificial makers on the surface of the tower. The environmental factors during the long‐term monitoring were discussed, especially the influence of temperature variation. Background modification was employed to reduce the error induced by the microchange of camera position. The long‐term 3‐D deformations of the ancient tower including vertical settlement and horizontal deformation in two directions during the shield tunneling construction were obtained. The results indicate that the proposed computer vision‐based method offers an efficient and cost‐effective approach to monitor the 3‐D deformation of the ancient structure and further for structural safety evaluation.

A Review of Computer Vision-Based Structural Deformation Monitoring in Field Environments

“…References [ 50 , 51 , 52 , 53 , 54 , 55 ] build S according to known physical dimensions on the surface of the object (such as the dimension of an artificial object or the dimension of the structural member obtained from the design drawing) and the corresponding image dimensions to measure the displacement of the structure.…”

“…Researchers began to use video analysis in civil engineering structure monitoring and achieved results. At present, computer vision-based monitoring technology has been applied to many fields of SHM, including bolt looseness detection [ 97 , 98 , 99 ]; quantifying disaster impact [ 100 , 101 , 102 , 103 , 104 ]; cable force monitoring [ 55 , 105 , 106 , 107 , 108 ]; modal frequency monitoring [ 2 , 19 , 52 , 63 , 74 , 95 ]; structural deformation measurement in 2D [ 34 , 109 , 110 , 111 , 112 ] and 3D [ 50 , 113 ]; bridge structural influence line [ 27 , 28 ] and influence surface measurement [ 29 ]; structural damage detection and localization [ 51 , 52 ]; updating finite element models of structures [ 114 , 115 ]; static and dynamic rotational angle measurement of large civil structures with high resolution [ 116 ], and spatio-temporal distribution of traffic loads [ 117 ].…”

“…When measuring a real structure outdoors, the position and direction of the camera often change slightly due to wind, vibration and ground instability [ 128 ]. Ye et al [ 50 ] believed that the thermal expansion and cold contraction of the mounting bracket would also cause a small change in the position of the camera, and this small motion of the camera, even a very small rotation, would lead to a very large error with increasing range. Figure 5 is a vision-based monitoring system in which the target is fixed, and Figure 5 a,b show the errors caused by camera translation and rotation, respectively.…”

“…Therefore, they are not suitable for long-term monitoring. BM and EC are relatively simple to calculate and have good adaptability to different scenarios, and researchers have carried out field experiments for up to four months to verify their effectiveness [ 50 , 132 ]. These are solutions that can be considered for widespread use.…”

A vision‐based surface displacement/strain measurement technique based on robust edge‐enhanced transform and algorithms for high spatial resolution

“…In recent years, self-sensing concrete or cement composite materials are under development. 10,11 Computer vision techniques have become popular in structural health monitoring [12][13][14][15][16] and damage detection [17][18][19] because they enable noncontact remote monitoring of structures with easy installation. For example, vision sensors have wide applications in one-point or target displacement monitoring based on key-point tracking, 1,20,21 digital image correlation (DIC), 22 or edge-enhanced information.…”