Characterization of Soil Crack Patterns Using Deep Neural Networks
Publication: Geo-Congress 2023
ABSTRACT
Desiccation cracks occur in fine-grained soils due to evaporation and volumetric shrinkage. Characterizing soil cracking patterns can offer new insights into the key controlling factors and the underlying soil desiccation cracking mechanisms. However, the formation of desiccation cracks inherently involves a complex network problem, posing a challenging task for modeling and monitoring purposes. This study employs a deep learning technique to develop a new framework for characterizing soil cracking. We utilize a deep learning (DL) based convolutional neural network (CNN) architecture to segment and characterize soil desiccation crack images gathered from peer-reviewed publications. Further, we present a methodology to evaluate the pattern identification capability by considering the overall metrics (precision, recall, and dice) and details (the number of crack segments, crack total length, average crack width, and surface crack ratio) of the soil crack features. The results of the proposed framework are compared with those attained from a classical image processing technique of binarization by thresholding. The comparison shows the presented DL-based method offers significant improvements in the performance and accuracy of pattern recognition.
Get full access to this article
View all available purchase options and get full access to this chapter.
REFERENCES
Cheng, Q., Tang, C. S., Zeng, H., Zhu, C., An, N., and Shi, B. (2020). Effects of microstructure on desiccation cracking of a compacted soil. Eng. Geol. (265)105418.
Cheng, Q., Tang, C. S., Xu, D., Zeng, H., and Shi, B. (2021). Water infiltration in a cracked soil considering effect of drying-wetting cycles. J. Hydrol. 593(125640).
Karimpouli, S., and Tahmasebi, P. (2019). Segmentation of digital rock images using deep convolutional autoencoder networks. Comput. Geosci. 126, 142–150.
Khan, S. H., He, X. M., Porikli, F., and Bennamoun, M. (2017). Forest change detection in incomplete satellite images with deep neural networks. IEEE Trans. Geosci. Remote Sens. 55, No. 9, 5407–5423.
Liu, W., Shi, H., He, X., Pan, S., Ye, Z., and Wang, Y. (2018). An application of optimized Otsu multi-threshold segmentation based on fireworks algorithm in cement SEM image. J. Algorithm. Comput. Technol. 13, 1–12.
Makasis, N., Narsilio, G. A., and Bidarmaghz, A. (2018). A machine learning approach to energy pile design. Comput. Geotech. 97, 189–203.
Mo, S. X., Zabaras, N., Shi, X. Q., and Wu, J. C. (2019). Deep autoregressive neural networks for high-dimensional inverse problems in groundwater contaminant source identification. Wat. Resour. Res. 55, No. 5, 3856–3881.
Otsu, N. (1979). A threshold selection method from gray-level histograms. IEEE Trans. Syst. Man Cybern. 9, No. 1, 62–66.
Ren, S., He, K., Girshick, R., and Sun, J. (2017). Faster R-CNN: towards real-time object detection with region proposal networks. IEEE Trans. Pattern Anal. Mach. Intell. 39, No. 6, 1137–1149.
Tang, C. S., Shi, B., Liu, C., Zhao, L. Z., and Wang, B. J. (2008). Influencing factors of geometrical structure of surface shrinkage cracks in clayey soils. Eng. Geol. 101, 204–217.
Tang, C. S., Cui, Y. J., Shi, B., Tang, A. M., and Liu, C. (2011). Desiccation and cracking behaviour of clay layer from slurry state under wetting-drying cycles. Geoderma 166 (1), 111–118.
Tang, C. S., Cheng, Q., Leng, T., Shi, B., Zeng, H., and Inyang, H. I. (2020). Effects of wetting- drying cycles and desiccation cracks on mechanical behavior of an unsaturated soil. CATENA 194, 104721.
Tang, C. S., Zhu, C., Cheng, Q., Zeng, H., Xu, J. J., Tian, B. G., and Shi, B. (2021). Desiccation cracking of soils: A review of investigation approaches, underlying mechanisms, and influencing factors. Earth-Science Reviews. 216, 103586.
Vahedifard, F., Jasim, F. H., Tracy, F. T., Abdollahi, M., Alborzi, A., and AghaKouchak, A. (2020). Levee Fragility Behavior under projected Future Flooding in a Warming climate. J. Geotech. Geoenviron. Eng. 146 (12), 04020139.
Xu, J. J., Zhang, H., Tang, C. S., Cheng, Q., Tian, B. G., Liu, B., and Shi, B. (2021). Automatic soil crack recognition under uneven illumination condition with the application of artificial intelligence. Eng. Geol. 296(106495).
Xu, J. J., Zhang, H., Tang, C. S., Cheng, Q., Liu, B., and Shi, B. (2022). Automatic soil desiccation crack recognition using deep learning. Geotechnique. 72(4) 337–349.
Zeng, H., Tang, C. S., Cheng, Q., Inyang, H. I., Rong, D. Z., Lin, L., and Shi, B. (2019). Coupling effects of interfacial friction and layer thickness on soil desiccation cracking behavior. Eng. Geol. 260(105220).
Zeng, H., Tang, C. S., Zhu, C., Vahedifard, F., Cheng, Q., and Shi, B. (2022). Desiccation cracking of soil subjected to different environmental relative humidity conditions. Eng. Geol. 297(106536).
Zhang, H., and Hong, X. G. (2019). Recent progresses on object detection: a brief review. Multimed. Tools Applic. 78, No. 19, 27809–27847.
Zhang, T. W., Deng, Y. F., Cui, Y. J., Lan, H. X., Zhang, F. Y., and Zhang, H. Y. (2019). Porewater salinity effect on flocculation and desiccation cracking behaviour of kaolin and bentonite considering working condition. Eng. Geol. 251, 11–23.
Information & Authors
Information
Published In
Geo-Congress 2023
Pages: 389 - 399
History
Published online: Mar 23, 2023
Authors
Metrics & Citations
Metrics
Citations
Download citation
If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download.