A Hybrid Machine Learning Framework for Construction Site Safety Assessment and Accident Risk Prediction in Infrastructure Projects
DOI:
https://doi.org/10.64882/ijrt.v14.i3.1826Keywords:
Construction safety · Accident risk prediction · Hybrid machine learning · Computer vision · Explainable AI · Infrastructure projects.Abstract
The status quo safety management approach is reactive, relying on the traditional, checklist-driven approach to safety in workplaces, including infrastructure projects like highways, bridges, tunnels and metro systems. This paper suggests a hybrid machine learning approach through combining three complementary information streams that are often treated separately: on-site observation of visual information, historical accident narratives and structured project risk factors. A vision stream identifies workers and personal protective equipment (PPE), detects proximity hazards, and derives site imagery based on employees' locations; a language stream extracts unstructured information about incidents and inspections from a free-text database for the purpose of latent causal factor mining; a structured stream models project, task and environmental attributes using gradient-boosted ensembles. A late-fusion reasoning layer is used to fuse the three streams into a single, interpretable site-level accident risk score, and a model-agnostic explainability module extracts the factors responsible for each score, which can then be used by site engineers to inform their actions. We explain the end-to-end architecture, an evaluation protocol for which construction safety data from the public is reported, and a series of baselines from the single-modality and classical approaches. Reported results are sample values intended to illustrate and not measured. The framework will be used as an early, explainable warning proactive decision support tool rather than a post incident analysis tool for infrastructure safety management.
References
Alsulamy, S., Alshayeb, M., Inam, I., Ahmed, A.: Machine learning applications for predicting safety incidents in the construction industry. Sci. Rep. 16 (2026). https://doi.org/10.1038/s41598-025-34763-0
Liu, W., Kang, X., Ye, Q., Xie, J.: Unravelling hierarchical penetration mechanisms and coupling relationships of safety risks in major transportation infrastructure construction using text mining and complex networks. Sci. Rep. 16, 7313 (2026). https://doi.org/10.1038/s41598-026-37778-3
Cheng, J.C.P., Wong, P.K.-Y., Luo, H., Wang, M., Leung, P.H.: Vision-based monitoring of site safety compliance based on worker re-identification and personal protective equipment classification. Autom. Constr. 139, 104312 (2022). https://doi.org/10.1016/j.autcon.2022.104312
Han, K., Zeng, X.: Deep learning-based workers' safety helmet wearing detection on construction sites using multi-scale features. IEEE Access 10, 718–729 (2022).
Hayat, A., Morgado-Dias, F.: Deep learning-based automatic safety helmet detection system for construction safety. Appl. Sci. 12(16), 8268 (2022). https://doi.org/10.3390/app12168268
Li, Y., Wei, H., Han, Z., Jiang, N., Wang, W., Huang, J.: Computer vision-based hazard identification of construction site using visual relationship detection and ontology. Buildings 12(6), 857 (2022). https://doi.org/10.3390/buildings12060857
Ding, Y., Liu, M., Luo, X.: Safety compliance checking of construction behaviors using visual question answering. Autom. Constr. 144, 104580 (2022). https://doi.org/10.1016/j.autcon.2022.104580
Gil, D., Lee, G.: Zero-shot monitoring of construction workers' personal protective equipment based on image captioning. Autom. Constr. 164, 105470 (2024). https://doi.org/10.1016/j.autcon.2024.105470
Chen, Z., Chen, H., Imani, M., Chen, R., Imani, F.: Vision language model for interpretable and fine-grained detection of safety compliance in diverse workplaces. Expert Syst. Appl. 265, 125769 (2025). https://doi.org/10.1016/j.eswa.2024.125769
Kim, H., Yi, J.S.: Image generation of hazardous situations in construction sites using a text-to-image generative model for training deep neural networks. Autom. Constr. 166, 105615 (2024).
Pan, X., et al.: Novel blockchain deep learning framework to ensure video security and lightweight storage for construction safety management. Adv. Eng. Inform. 59, 102334 (2024).
Wang, Y., Xiao, B., Bouferguene, A., Al-Hussein, M.: Proactive safety hazard identification using visual–text semantic similarity for construction safety management. Autom. Constr. 166, 105602 (2024).
Zhu, Y., Zhou, J., Zhang, B., Wang, H., Huang, M.: Statistical analysis of major tunnel construction accidents in China from 2010 to 2020. Tunn. Undergr. Space Technol. 124, 104460 (2022).
Tang, C., Shen, C., Zhang, J., Guo, Z.: Identification of safety risk factors in metro shield construction. Buildings 14, 492 (2024).
Huo, X., Du, S., Jiao, L.: Critical causal path analysis of subway construction safety accidents based on text mining. ASCE-ASME J. Risk Uncertain. Eng. Syst. Part A Civ. Eng. 11, 04024075 (2025).
Su, G.: Research on cause analysis and management of coal mine safety risk based on social network and bow-tie model. Sci. Rep. 15, 29850 (2025).
Chen, T., Guestrin, C.: XGBoost: a scalable tree boosting system. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp. 785–794 (2016). https://doi.org/10.1145/2939672.2939785
Lundberg, S.M., Lee, S.-I.: A unified approach to interpreting model predictions. In: Advances in Neural Information Processing Systems 30, pp. 4765–4774 (2017).
Devlin, J., Chang, M.-W., Lee, K., Toutanova, K.: BERT: pre-training of deep bidirectional transformers for language understanding. In: Proceedings of NAACL-HLT, pp. 4171–4186 (2019).
Downloads
How to Cite
Issue
Section
License
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
