An Urban Road Risk Assessment Framework Based on Convolutional Neural Networks

Abstract

In contemporary cities, road collapse is one of the most common disasters. This study proposed a framework for assessing the risk of urban road collapse. The framework first established a risk indicator system that combined environmental and anthropogenic factors, such as soil type, pipeline, and construction, as well as other indicators. Second, an oversampling technique was used to create the dataset. The framework then constructed and trained a convolutional neural network (CNN)-based model for risk assessment. The experimental results show that the CNN model (accuracy: 0.97, average recall: 0.91) outperformed other models. The indicator contribution analysis revealed that the distance between the road and the construction site (contribution: 0.132) and the size of the construction (contribution: 0.144) are the most significant factors contributing to road collapse. According to the natural breaks, a road collapse risk map of Foshan City, Guangdong Province, was created, and the risk level was divided into five categories. Nearly 3% of the roads in the study area are at very high risk, and 6% are at high risk levels, with the high risk roads concentrated in the east and southeast. The risk map produced by this study can be utilized by local authorities and policymakers to help maintain road safety.

Keywords

Convolutional neural networks / Data augmentation / Risk assessment / Urban road collapse

References

[1] Chen JL, Huang GR, Chen WJ. Towards better flood risk management: Assessing flood risk and investigating the potential mechanism based on machine learning models. Journal of Environmental Management, 2021, 293: 112810

[2] China Meteorological Administration. 2022. Guangdong: Pursue advantages and avoid disadvantages to deal with typhoon "Nige". http://www.cma.gov.cn/2011xwzx/2011xgzdt/202211/t20221107_5166943.html. Accessed 15 May 2023 (in Chinese).

[3] Cui L, Huang J, Wu X, Chen J, Jiang G, Xu S. The control of the formation of karst collapse through the ancient Yellow River Fault zone in Xuzhou China. Carbonates and Evaporites, 2017, 34(1): 1-10

[4] Department of Emergency Management of Guangdong. 2019. Investigation report on the major accident of the collapse of rail transit line 2 in Foshan City, Guangdong Province, 2018. http://yjgl.gd.gov.cn/gk/zdlyxxgk/sgdcbg/content/post_2511537.html. Accessed 15 May 2023 (in Chinese).

[5] Department of Emergency Management of Guangzhou. 2020. Guangzhou Tianhe District under construction rail transit collapse accident investigation report. http://yjglj.gz.gov.cn/zwgk/zdlyxxgk/scaqsgdcbgxx/content/post_5832879.html. Accessed 15 May 2023 (in Chinese).

[6] Dong ZH, Zhang RX, Shao XL, Kuang ZS. Learning sparse features with lightweight ScatterNet for small sample training. Knowledge-Based Systems, 2020, 205: 106315

[7] Foshan Bureau of Transportation. 2021. 2019 traffic statistics. http://jtys.foshan.gov.cn/zwgk/tjsj/content/post_4868689.html. Accessed 15 May 2023 (in Chinese).

[8] Gao L, Lu P, Ren YH. A deep learning approach for imbalanced crash data in predicting highway-rail grade crossings accidents. Reliability Engineering & System Safety, 2021, 216: 108019

[9] Gao X, Chen T, Niu RQ, Plaza A. Recognition and mapping of landslide using a fully convolutional DenseNet and influencing factors. IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, 2021, 14: 7881-7894

[10] Hong DF, Gao LR, Yokoya N, Yao J, Chanussot J, Du Q, Zhang B. More diverse means better: Multimodal deep learning meets remote-sensing imagery classification. IEEE Transactions on Geoscience and Remote Sensing, 2021, 59(5): 4340-4354

[11] Huang FM, Huang JS, Jiang SH, Zhou CB. Landslide displacement prediction based on multivariate chaotic model and extreme learning machine. Engineering Geology, 2017, 218: 173-186

[12] Jena R, Pradhan B, Beydoun G, Alamri AM, Ardiansyah Nizamuddin, Sofyan H. Earthquake hazard and risk assessment using machine learning approaches at Palu Indonesia. Science of the Total Environment, 2020, 749: 141582

[13] Jena R, Pradhan B, Naik SP, Alamri AM. Earthquake risk assessment in NE India using deep learning and geospatial analysis. Geoscience Frontiers, 2021, 12(3): 101110

[14] Jiang, T., S.K. Bendre, H. Lyu, and J. Luo. 2021. From static to dynamic prediction: Wildfire risk assessment based on multiple environmental factors. In Proceedings of the 2021 IEEE International Conference on Big Data (Big Data), 15–18 December 2021, Orlando, FL, USA, 4877–4886.

[15] Katrakazas C, Quddus M, Chen WH. A new integrated collision risk assessment methodology for autonomous vehicles. Accident Analysis and Prevention, 2019, 127: 61-79

[16] Kiani J, Camp C, Pezeshk S. On the application of machine learning techniques to derive seismic fragility curves. Computers & Structures, 2019, 218: 108-122

[17] Kourehpaz P, Hutt CM. Machine learning for enhanced regional seismic risk assessments. Journal of Structural Engineering, 2022

[18] Krizhevsky A, Sutskever I, Hinton GE. ImageNet classification with deep convolutional neural networks. Communications of the ACM, 2017, 60(6): 84-90

[19] Kuliczkowska E. The interaction between road traffic safety and the condition of sewers laid under roads. Transportation Research Part D: Transport and Environment, 2016, 48: 203-213

[20] LeCun Y, Boser B, Denker JS, Henderson D, Howard RE, Hubbard W, Jackel LD. Backpropagation applied to handwritten zip code recognition. Neural Computation, 1989, 1(4): 541-551

[21] Li YT, Hong HY. Modelling flood susceptibility based on deep learning coupling with ensemble learning models. Journal of Environmental Management, 2023, 325: 116450

[22] Li YY, Sheng YF, Chai B, Zhang W, Zhang TL, Wang JJ. Collapse susceptibility assessment using a support vector machine compared with back-propagation and radial basis function neural networks. Geomatics Natural Hazards & Risk, 2020, 11(1): 510-534

[23] Li WB, Fan XM, Huang FM, Chen W, Hong HY, Huang JS, Guo ZZ. Uncertainties analysis of collapse susceptibility prediction based on remote sensing and GIS: Influences of different data-based models and connections between collapses and environmental factors. Remote Sensing, 2020, 12(24): 4134

[24] Liu Y, Tan F, Jiao Y, Wang J, Wang W. Risk assessment of ground collapse based on finite interval cloud model. Safety and Environmental Engineering, 2021, 28(4): 115-120.

[25] Lv L, Chen T, Dou J, Plaza A. A hybrid ensemble-based deep-learning framework for landslide susceptibility mapping. International Journal of Applied Earth Observation and Geoinformation, 2022, 108: 102713

[26] Ministry of Transport of the People's Republic of China. 2021. 2020 national highway statistics bulletin. https://xxgk.mot.gov.cn/2020/jigou/glj/202110/t20211027_3623195.html. Accessed 15 May 2023 (in Chinese).

[27] National Weather Science Data Center. 2022. Hourly observation data of surface meteorological stations in China. https://data.cma.cn/data/cdcdetail/dataCode/A.0012.0001.html. Accessed 15 May 2023 (in Chinese).

[28] Papadopoulou-Vrynioti K, Bathrellos GD, Skilodimou HD, Kaviris G, Makropoulos K. Karst collapse susceptibility mapping considering peak ground acceleration in a rapidly growing urban area. Engineering Geology, 2013, 158: 77-88

[29] Shi H, Chen Y, Chen X. Summary of research on SMOTE oversampling and its improved algorithms. CAAI Transactions on Intelligent Systems, 2019, 14: 1073-1083.

[30] Wang X, Xu Y. Investigation on the phenomena and influence factors of urban ground collapse in China. Natural Hazards, 2022, 113(1): 1-33

[31] Wang K, Zhang JW, Gao GQ, Qiu JL, Zhong YJ, Guo CX, Zhao WC, Tang KJ, Su XL. Causes, risk analysis, and countermeasures of urban road collapse in China from 2019 to 2020. Journal of Performance of Constructed Facilities, 2022

[32] Wei AH, Li D, Zhou YH, Deng Q, Yan LD. A novel combination approach for karst collapse susceptibility assessment using the analytic hierarchy process, catastrophe, and entropy model. Natural Hazards, 2021, 105(1): 405-430

[33] Xiao W, Tian WP. Hazard assessment and zoning of collapse along highways in China based on backward cloud algorithm. Geomatics Natural Hazards & Risk, 2019, 10(1): 1227-1241

[34] Yan, J., Y. Cheng, A. Zhang, and Y. Zhao. 2012. The evaluation of highway collapse hazard based on the fuzzy comprehensive evaluation method. In Proceedings of the 2nd International Conference on Civil Engineering and Transportation (ICCET 2012), 27–28 October 2013, Guilin, China, 1593–1597.

[35] Yao X. Evolving artificial neural networks. Proceedings of the IEEE, 1999, 87(9): 1423-1447

[36] Yilmaz I, Marschalko M, Bednarik M. An assessment on the use of bivariate, multivariate and soft computing techniques for collapse susceptibility in GIS environ. Journal of Earth System Science, 2013, 122(2): 371-388

[37] Zhai J, Qi J, Zhang S. Imbalanced data classification based on diverse sample generation and classifier fusion. International Journal of Machine Learning and Cybernetics, 2021, 13(3): 735-750

[38] Zhang Y, Zhu W, Zhao C, Han B. Moniting and inversion of Foshan metro collapse with multi-temporal Insar and field investigation. Journal of Engineering Geology, 2021, 29(4): 1167-1177.

[39] Zhang A, Yu H, Huan Z, Yang X, Zheng S, Gao S. SMOTE-RkNN: A hybrid re-sampling method based on SMOTE and reverse k-nearest neighbors. Information Sciences, 2022, 595: 70-88

[40] Zhao W, Gan FP, Meng Y, Zheng ZJ, Liu XM. Application of seismic velocity tomography in investigation of karst collapse hazards, Guangzhou. China. Environmental Earth Sciences, 2018