Structural health monitoring (SHM) is essential for ensuring the safety and durability of large structures by identifying and mitigating damage risks. Computer vision (CV)-based measurement systems offer advantages over conventional sensors, including broader coverage of inaccessible areas and lower installation costs. This work introduces a novel noncontact CV-based physics-guided framework for measuring structural vibration responses under ambient (stochastic) conditions from its vibration videos, localizing damages associated with mass and stiffness changes, and quantifying their severity. Homography transformation and integration of the speeded-up robust features and Kanade–Lucas–Tomasi tracking algorithms are employed to extract displacement response from vibration video under different camera orientations. Unlike conventional approaches, this study employs principal component analysis as an unsupervised data-driven approach to estimate the optimal system order from displacement responses. State-space matrices’ estimation and potential damage identification are performed using the numerical algorithms for subspace state-space system identification and stochastic damage locating vector approach, respectively. An improved Bayesian inference-based stochastic model updating algorithm is proposed to precisely locate the actual damaged elements of the structure and assess their severity. The proposed methodology is validated on a numerical truss model with simulated damage scenarios and a laboratory-scale wooden truss bridge using sensor-recorded data. The methodology is finally validated on the damage localization for a photo-realistic synthetic truss bridge model from its vibration videos. The improved accuracy of damage severity prediction compared to existing techniques demonstrates the method’s adeptness in precisely detecting structural damage. Furthermore, the methodology showcases its efficacy in locating damage across various severity levels, facilitating its application in real-time SHM.
FanWQiaoP.Vibration-based damage identification methods: a review and comparative study. Struct Health Monit2011; 10(1): 83–111.
2.
AshoryM. Correction of mass-loading effects of transducers and suspension effects in modal testing. In: Proceedings of the 16th international modal analysis conference, Santa Barbara CA, 2–5 February 1998, p. 3243. Bethel CT, United States: Society for Experimental Mechanics.
3.
FarrarCRDarlingTWMiglioriA, et al. Microwave interferometers for non-contact vibration measurements on large structures. Mech Syst Signal Process1999; 13(2): 241–253.
4.
KassotakisNSarhosisV.Employing non-contact sensing techniques for improving efficiency and automation in numerical modelling of existing masonry structures: a critical literature review. Structures2021; 32: 1777–1797.
5.
SaravananTJGopalakrishnanNRaoNP.Damage detection in structural elements through propagating waves using radially weighted and factored RMS. Measurement2015; 73: 520–538.
6.
LucasBDKanadeT.An iterative image registration technique with an application to stereo vision. In: Proceedings of the DARPA image understanding workshop, Vancouver, 24-28 August 1981, pp. 121-130. Burlington, MA: Morgan Kaufmann Publishers Inc.
7.
TomasiCKanadeT.Tracking of point features. Int J Comput Vis1991; 9: 137–154.
8.
TomasiCShiJ. Direction of heading from image deformations. In: Proceedings of IEEE conference on computer vision and pattern recognition, New York, NY, 15-17 June 1993, pp. 422–427. Piscataway, New Jersey: IEEE.
9.
WadhwaNRubinsteinMDurandF, et al. Phase-based video motion processing. ACM Trans Graph2013; 32(4): 1–10.
10.
ChenJGWadhwaNChaYJ, et al. Modal identification of simple structures with high-speed video using motion magnification. J Sound Vib2015; 345: 58–71.
11.
SarrafiAMaoZNiezreckiC, et al. Vibration-based damage detection in wind turbine blades using phase-based motion estimation and motion magnification. J Sound Vib2018; 421: 300–318.
12.
ChenJGAdamsTMSunH, et al. Camera-based vibration measurement of the World War I memorial bridge in Portsmouth, New Hampshire. J Struct Eng2018; 144(11): 1–19.
13.
SaravananTJNishioMSuzukiT.Computer vision-based in-plane strain field monitoring for identification of local damages in structural members. Struct Health Monit2019; 2019: 32465.
14.
CollierSDareT.Regime sorting for multiscale vibrations and phase-based motion extraction. In Rotating machinery, optical methods & scanning LDV methods, volume 6: proceedings of the 40th IMAC, a conference and exposition on structural dynamics, Daytona Beach, FL, 31 Jan 31–3 Feb, 2022, July 2022, pp. 61–71. Cham: Springer International Publishing.
15.
SaravananTJGopalakrishnanNHariBK.Damage identification in structural elements through curvature mode shapes and nonlinear energy operator. Compos Mater Eng2019; 1(1): 33–48.
16.
ZhaiGNarazakiYWangS, et al. Synthetic data augmentation for pixel-wise steel fatigue crack identification using fully convolutional networks. Smart Struct Syst2022; 29(1): 237–250.
17.
YaoZJiangSWangS, et al. Intelligent crack identification method for high-rise buildings aided by synthetic environments. Struct Des Tall Spec Build2024; 33: e2117.
18.
DangHVRazaMNguyenTV, et al. Deep learning-based detection of structural damage using time-series data. Struct Infrastruct Eng2021; 17(11): 1474–1493.
19.
NarazakiYHoskereVYoshidaK, et al. Synthetic environments for vision-based structural condition assessment of Japanese high-speed railway viaducts. Mech Syst Signal Process2021; 160: 107850.
20.
NarazakiYGomezFHoskereV, et al. Efficient development of vision-based dense three-dimensional displacement measurement algorithms using physics-based graphics models. Struct Health Monit2021; 20(4): 1841–1863.
21.
SohnHFarrarCR.Damage diagnosis using time series analysis of vibration signals. Smart Mater Struct2001; 10(3): 446.
22.
FassoisSDSakellariouJS.Time-series methods for fault detection and identification in vibrating structures. Philos Trans R Soc A Math Phys Eng Sci2007; 365(1851): 411–448.
23.
ThiyagarajanJSSiringoringoDMWangchukS, et al. Implementation of video motion magnification technique for non-contact operational modal analysis of light poles. Smart Struct Syst2021; 27: 227.
24.
LakshmiKRaoARMGopalakrishnanN.Singular spectrum analysis combined with ARMAX model for structural damage detection. Struct Control Health Monit2017; 24(9): e1960.
25.
BhowmickBTripuraTHazraB.Real-time structural modal identification using recursive canonical correlation analysis and application towards online structural damage detection. J Sound Vib2019; 468: 115101.
26.
BernalD.Extracting flexibility matrices from state-space realizations. In COST F3 conference, Madrid, Spain, 6–9 June, 2000, vol. 1, 2000, pp. 127135. Madrid, Spain: Universidad Politécnica de Madrid.
27.
BernalD.Load vectors for damage localization. J Eng Mech2002; 128(1): 7–14.
28.
BernalD.Flexibility-based damage localization from stochastic realization results. J Eng Mech2006; 132(6): 651–658.
29.
BernalD.Damage localization from the null space of changes in the transfer matrix. AIAA J2007; 45(2): 374–381.
30.
BernalD.Load vectors for damage location in systems identified from operational loads. J Eng Mech2010; 136(1): 31–39.
31.
MarinLDöhlerMBernalD, et al. Robust statistical damage localization with stochastic load vectors. Struct Control Health Monit2015; 22(3): 557–573.
32.
ChangCMChouJY.Deformation estimation of truss bridges using two-stage optimization from cameras. Smart Struct Syst2023; 31(4): 409–419.
33.
TanDDingZLiJ, et al. Target-free vision-based approach for vibration measurement and damage identification of truss bridges. Smart Struct Syst2023; 31(4): 421–436.
34.
NiYQLiuWQRuiEZ, et al. A novel computer vision-based vibration measurement and coarse-to-fine damage assessment method for truss bridges. Smart Struct Syst2023; 31(4): 393–407.
35.
DangHVRazaMNguyenTV, et al. Deep learning-based detection of structural damage using time-series data. Struct Infrastruct Eng2021; 17(11): 1474–1493.
36.
BaiYZhaBSezenH, et al. Engineering deep learning methods on automatic detection of damage in infrastructure due to extreme events. Struct Health Monit2023; 22(1): 338–352.
37.
DuranBEftekhar AzamSSanayeiM.Leveraging deep learning for robust structural damage detection and classification: a transfer learning approach via CNN. Infrastructures2024; 9(12): 229.
38.
LoopCZhangZ. Computing rectifying homographies for stereo vision. In: Proceedings of 1999 IEEE computer society conference on computer vision and pattern recognition, 1999, p. 1.
39.
BayHTuytelaarsTVan GoolL.SURF: speeded up robust features. In: LeonardisABischofHPinzA (eds.) Computer vision – ECCV 2006, ECCV 2006, Lecture notes in computer science, vol. 3951. pp. 404–417. Berlin, Heidelberg: Springer, 2006.
40.
HarrisCStephensM. A combined corner and edge detector. In: Proceedings of the 4th Alvey vision conference, Manchester, UK, September 1988. Manchester, UK: Alvey Vision Club.
41.
BirchfieldS. Derivation of Kanade-Lucas-Tomasi tracking equation. Unpublished notes, 1997.
42.
Van OverscheePDe MoorB. N4SID: subspace algorithms for the identification of combined deterministic-stochastic systems. Automatica1994; 30(1): 75–93.
43.
ShokraviHShokraviHBakharyN, et al. Health monitoring of civil infrastructures by subspace system identification method: an overview. Appl Sci2020; 10(8): 2786.
44.
KaushikASundaramBAParivallalS, et al. Numerical studies on the damage locating vector approach for damage detection using static load data. Procedia Eng2014; 86: 693–700.
45.
WanHPRenWX.Stochastic model updating utilizing Bayesian approach and Gaussian process model. Mech Syst Signal Process2016; 70: 245–268.
46.
IsabonaJImoizeALOjoS, et al. Machine learning-based GPR with LBFGS Kernel parameters selection for optimal throughput mining in 5G wireless networks. Sustainability2023; 15(2): 1678.
47.
BeckJLAuSK.Bayesian updating of structural models and reliability using Markov chain Monte Carlo simulation. J Eng Mech2002; 128(4): 380–391.
48.
LiPJXuDWZhangJ.Probability-based structural health monitoring through Markov chain Monte Carlo sampling. Int J Struct Stab Dyn2016; 16(07): 1550039.
49.
ZhangLDongLChengS, et al. Efficient reliability assessment method for bridges based on Markov Chain Monte Carlo (MCMC) with Metropolis-Hasting Algorithm (MHA). In: IOP conference series: earth and environmental science, vol. 580, no. 1, October 2020, p. 012030. Bristol, United Kingdom: IOP Publishing Ltd.
50.
BabaeiMJalilkhaniMMollaeiS.A numerical method for estimating the dynamic response of structures. J Civ Environ Eng2021; 29(1): 1–19.
51.
HewettDR.A fast non-linear, finite element solver for earthquake response of buildings. Thesis, University of Canterbury, Christchurch, New Zealand, 2010.
52.
BanerjeeSSaravananTJ.Utilizing data-driven algorithms for blind modal parameter identification of structures from output-only video measurements. Structures2024; 63: 106410.
53.
KullaaJ.Sequential structural health monitoring and damage detection. In: BollerC (ed.) Proceedings of the 6th European workshop on structural health monitoring. Dresden, Germany: DGZfP, 2012, pp. 674–681.
54.
KullaaJ.Distinguishing between sensor fault, structural damage, and environmental or operational effects in structural health monitoring. Mech Syst Signal Process2011; 25(8): 2976–2989.
55.
KullaaJ. Sensor validation using minimum mean square error estimation. Mechanical Systems and Signal Processing2010; 24(5): 1444–1457.
