Abstract
Civil structures like buildings, bridges, roads, pavements, tunnels and masonry buildings need frequent
monitoring to ensure structural safety and serviceability. The traditional inspection methods are mainly based on
visual inspections by a human; which are generally time-consuming, subjective and challenging on large-scale
infrastructure. While efforts over the past few years using deep learning have substantially enhanced automated
crack detection, there is currently limited support for engineering assessment and maintenance planning for
cracks detected on the surface, which are present in many existing crack detection systems. This paper
introduces the Autonomous Structural Health Monitor (ASHM), an intelligent multi-modal software technology
platform that assists structural engineers to monitor infrastructure through deep learning and computer vision. It
supports three operations: Batch Image Inspection, Static Camera Inspection and Live Infrastructure Inspection,
which allows an inspection in various practical scenarios. The AI engine combines Deep ResNet50 (Crack
classification) and U-Net (Pixel-level crack segmentation). The segmented regions of the crack are further
processed using OpenCV contours, pixel count and skeletonization to try and estimate area of crack, length of
crack and maximum width of crack. The information from the measurements is taken into account by a rulebased module to assess and classify the damage and its grade, and to provide maintenance recommendations.
Lastly, ASHM generates a professional PDF inspection report automatically with the identified properties of the
cracks, structural assessment and recommended measures for maintenance works. A set of images of concrete
cracks (Concrete Crack Images, CFD) and an in-house image database were selected for experimental
evaluation combined with SDNET2018. The results achieved prove the reliability of the proposed platform that
combines practical inspection workflows in a single software solution, while detecting cracks. ASHM's modular
structure also opens the door to possibility of future integration with a variety of other IoT devices, as well as
drones and physical, non-destructive testing methods for advanced, structural health monitoring. This work
employs a Deep Learning network (ResNet50), a super pixel segmentation algorithm known as U-Net,
Computer Vision techniques for crack detection and OpenCV software for further report generation to carry out
Structural Health Monitoring of civil structures.
References
1) ]K. X. Zhang, S. Ren, J. Sun, and He, "Deep Residual Learning for Image Recognition," Proceedings
of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Las Vegas, NV, USA,
pp. 770–778, 2016.
2) O. Ronneberger, P. Fischer and T. Brox, "U-Net: Convolutional Networks for Biomedical Image
Segmentation," Medical Image Computing and Computer-Assisted Intervention (MICCAI), LNCS
9351, Springer, pp. 234–241, 2015.
3) S. Dorafshan, R. J. Thomas and M. Maguire, “SDNET2018: An Annotated Image Dataset for NonContact Concrete Crack Detection Using Deep Convolutional Neural Networks”, Data in Brief, vol. 21,
pp. 1664–1668, 2018.
4) S. Dorafshan and M. Maguire, "Autonomous Detection of Concrete Cracks on Bridge Decks Using
Deep Learning," Proceedings of Digital Imaging, 2017.
5) A. Zhang, K. C. P. Wang, B. Li, E. Yang, X. Dai, Y. Peng and others, "Automated Pixel-Level
Pavement Crack Detection on 3D Asphalt Surfaces Using a Deep-Learning Network", Computer-Aided
Civil and Infrastructure Engineering, vol. 32, no. 10, pp. 805–819, 2017.
6) J. Long, E. Shelhamer, and T. Darrell, “Fully Convolutional Networks for Semantic Segmentation,” in
Proceedings of CVPR, pp. 3431–3440, 2015.
7) G. Bradski, "The OpenCV Library," Dr. Dobb's Journal of Software Tools, 2000.
8) D. P. Kingma, J. Ba: Adam: A Method for Stochastic Optimization, International Conference on
Learning Representations (ICLR), 2015.
9) I. Goodfellow, Y. Bengio and A. Courville, Deep Learning, MIT Press, Cambridge, MA, USA,
2016The Manila Times (2018, April 26). At a glance: The Philippine Health Care System
10) R. Szeliski, Computer Vision: Algorithms and Applications, 2nd Ed., Springer, 2022.
11) S. Shomal Zadeh, S. Aalipour Birgani, M. Khorshidi, and F. Kooban, “Concrete Surface Crack
Detection with Convolutional Based Deep Learning Models,” 2024
12) R. C. Gonzalez and R. E. Woods, Digital Image Processing, 4th ed., Pearson, 2018.
13) Y. LeCun, Y. Bengio, and G. Hinton, "Deep Learning," Nature, vol. 521, no. 7553, pp. 436–444, 2015.
14) TensorFlow Developers, TensorFlow: Large-Scale Machine Learning on Heterogeneous Systems,
Google, 2015.
15) F. Pedregosa et al., "Scikit-learn: Machine Learning in Python," Journal of Machine Learning
Research, vol. 12, pp. 2825–2830, 2011.