Development and Demonstration of a Deep Learning-Based AI CCTV Control Model for SIF Prevention in Power Generation Companies

doi:10.15683/kosdi.2025.12.31.975

All Issue

2025 Vol.21, Issue 4 Preview Page Next Page

Research Article

Development and Demonstration of a Deep Learning-Based AI CCTV Control Model for SIF Prevention in Power Generation Companies 발전사 중대재해 예방을 위한 딥러닝 기반 AI CCTV 관제모델 개발 및 실증 연구: 이창덕¹ㆍ장관수²ㆍ김홍석³ㆍ손지영⁴ㆍ임성윤⁵ㆍ이준원⁶*
Chang-Deok Lee¹, Gwan-Su Jang², Hong-Seok Kim³, Ji-Young Son⁴, Sung-Yun Lim⁵, Joon-Won Lee⁶*; ¹Doctor’s Course, Department of Safety & Health Convergence Engineering, Soongsil University, Seoul, Republic of Korea
²Ph.D, Department of Safety & Health Convergence Engineering, Soongsil University, Seoul, Republic of Korea
³Doctor’s Course, Department of Safety & Health Convergence Engineering, Soongsil University, Seoul, Republic of Korea
⁴Doctor’s Course, Department of Safety & Health Convergence Engineering, Soongsil University, Seoul, Republic of Korea
⁵Doctor’s Course, Department of Safety & Health Convergence Engineering, Soongsil University, Seoul, Republic of Korea
⁶Professor, Department of Safety & Health Convergence Engineering, Soongsil University, Seoul, Republic of Korea

31 December 2025. pp. 975-983

PDF

Abstract

Purpose: This study aims to empirically verify the effectiveness of a deep learning–based AI CCTV control system in preventing recurring serious industrial accidents(SIFs: falls, entrapments, and collisions) at power plants. Method: Industrial accident data from 2018–2022 and AI CCTV detection performance data from 2023–2025 were analyzed. Descriptive statistics, trend analysis, Difference-in-Differences(DiD), and event study models were applied to assess pre- and post-adoption effects. Result: SIFs accounted for 46.8% of all industrial accidents. After AI CCTV implementation, helmet detection improved from 23.78% to 97.56%, and fall detection from 37.79% to 87.79%. The DiD analysis showed a significant adoption effect of +0.368(p<0.01). Event study analysis further confirmed that the preventive effects gradually strengthened and became institutionalized. Conclusion: By integrating accident statistics with AI performance data, this study quantitatively demonstrated the accident prevention effect of AI CCTV. The system overcame the limitations of human-centered monitoring and shifted safety management from reactive response to proactive prevention. This smart safety management system will help prevent serious accidents in high-risk sites.

Keywords

Serious Injuries and Fatalities(SIF)

Accident Prevention

Deep Learning

AI CCTV

Power Plants

연구목적: 본 연구는 발전소에서 반복 발생하는 중대재해(SIF: 떨어짐, 끼임, 부딪힘) 예방을 위해 딥러닝 기반 AI CCTV 관제시스템을 구축하고 효과를 실증적으로 검증하는데 있다. 연구방법: 2018~2022년 안전사고 데이터와 2023~2025년 탐지율 성능데이터를 활용하여 기술통계, 추세분석, 이중차분(DiD), 이벤트 스터디 모형으로 AI도입 전후 효과를 분석한다. 연구결과: 중대재해는 전체 사고의 46.8%를 차지하였으며, 도입부서의 안전모 탐지율은 23.78%→97.56%, 쓰러짐은 37.79%→87.79%로 향상되었다. DiD 분석결과, AI 도입 효과( +0.368(p<0.01)가 확인되었고, 이벤트 스터디 분석에서도 효과가 점차 강화되고 내재화됨을 입증하였다. 결론: 안전사고 통계와 AI 성능데이터 결합을 통해 사고예방효과를 계량적으로 입증하였고, AI CCTV 감시체계는 인력중심 감시한계를 극복하고 안전관리를 사후 대응에서 사전 예방중심으로 전환해야함을 보여준다. 이러한 스마트 안전관리체계는 고위험 현장의 중대재해 예방에 기여할 것이다.

키워드

중대재해(SIF)

산업재해 예방

딥러닝

AI CCTV

발전소

References

ILO (2020). Safety and Health at the Heart of the Future of Work. International Labour Organization, Geneva.

OSHA (2021). Leading Indicators for Safety and Health Programs. U.S. Department of Labor, Washington, DC.

Redmon, J., arhadi, A. (2018). “YOLOv3: An incremental improvement.” arXiv preprint arXiv:1804.02767.

S Power Plant (2023). Safety Accident Analysis Data.

S Power Plant (2025). AI CCTV Control System Construction Data.

Wang, C.Y., Bochkovskiy, A., Liao, H.Y.M. (2023). “YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors.” CVPR Workshops, Vancouver, Canada, pp. 746-755. 10.1109/CVPR52729.2023.00721

Yan, S., Xiong, Y., Lin, D. (2018). “Spatial temporal graph convolutional networks for skeleton-based action recognition.” AAAI Conference on Artificial Intelligence, New Orleans, Louisiana, USA, Vol. 32, No. 1, pp. 7444-7452 10.1609/aaai.v32i1.12328

Zeng, W., Zheng, C., Zhu, S., Mendieta, M., Yang, T., Chen, C., Ding, Z. (2021). “PoseFormer: Direct pose forecasting with Transformers.” ICCV 2021, Virtual, pp. 11656-11665.

Information

Publisher :The Korean Society of Disaster Information
Publisher(Ko) :한국재난정보학회
Journal Title :Journal of the Society of Disaster Information
Journal Title(Ko) :한국재난정보학회논문집
Volume : 21
No :4
Pages :975-983
DOI :https://doi.org/10.15683/kosdi.2025.12.31.975

[1] ILO (2020). Safety and Health at the Heart of the Future of Work. International Labour Organization, Geneva.

[2] OSHA (2021). Leading Indicators for Safety and Health Programs. U.S. Department of Labor, Washington, DC.

[3] Redmon, J., arhadi, A. (2018). “YOLOv3: An incremental improvement.” arXiv preprint arXiv:1804.02767.

[4] S Power Plant (2023). Safety Accident Analysis Data.

[5] S Power Plant (2025). AI CCTV Control System Construction Data.

[6] Wang, C.Y., Bochkovskiy, A., Liao, H.Y.M. (2023). “YOLOv7: Trainable bag-of-freebies sets new state-of-the-art for real-time object detectors.” CVPR Workshops, Vancouver, Canada, pp. 746-755. 10.1109/CVPR52729.2023.00721

[7] Yan, S., Xiong, Y., Lin, D. (2018). “Spatial temporal graph convolutional networks for skeleton-based action recognition.” AAAI Conference on Artificial Intelligence, New Orleans, Louisiana, USA, Vol. 32, No. 1, pp. 7444-7452 10.1609/aaai.v32i1.12328

[8] Zeng, W., Zheng, C., Zhu, S., Mendieta, M., Yang, T., Chen, C., Ding, Z. (2021). “PoseFormer: Direct pose forecasting with Transformers.” ICCV 2021, Virtual, pp. 11656-11665.

Journal of the Society of Disaster Information ISSN:1976-2208(Print) 2671-5287(Online) 한국재난정보학회논문집

All Issue