Helmet Detection System with YOLO Architecture in Real-Time Construction Site Cameras

Abdil Karakan; Yüksel Oğuz

doi:10.54365/adyumbd.1757010

TR EN

Gerçek Zamanlı Şantiye Kameralarından YOLO Mimarisi ile Otomatik Kask Algılama Sistemi

Öz

İnşaat alanlarında iş kazalarının önemli bir kısmı yukarıdan düşen nesnelerden kaynaklanmaktadır. Bu nedenle kask kullanımı, çalışanların güvenliği için kritik öneme sahiptir. Ancak kask kontrolü genellikle ekip başları tarafından gözlem yoluyla yapılmakta ve insan hatasından dolayı eksiklikler yaşanmaktadır. Bu çalışmada, şantiye kameralarından elde edilen görüntülerle gerçek zamanlı otomatik kask tespiti gerçekleştirilmiştir. Gerçek ortamdan 4.759 görüntü toplanmış, farklı açılar, mesafeler ve ışık koşulları göz önünde bulundurulmuştur. Veri artırma yöntemleri (gri tonlama, bulanıklaştırma, gürültü ekleme, parlaklık ve renk doygunluğu değişimi, perspektif dönüşümü, yeniden boyutlandırma vb.) uygulanarak veri sayısı 17.781’e çıkarılmıştır. Böylece veri seti gerçek ortama uyarlanmıştır. Elde edilen veri seti YOLO’nun dört farklı sürümünde test edilmiş, en yüksek başarı YOLOv8 ile sağlanmıştır. Bu modelde doğruluk %98.3, F1-Skoru 0.939 ve mAP@.5 değeri 0.977 olarak bulunmuş, FPS değeri 27’den 60’a yükselmiştir.

Anahtar Kelimeler

Kask tespiti, Kişisel koruma ekipmanı, Derin öğrenme, Nesne tespiti, YOLO mimarisi

Helmet Detection System with YOLO Architecture in Real-Time Construction Site Cameras

Öz

A significant proportion of construction site accidents occur due to falling objects from above, making helmet usage critical for worker safety. Helmet compliance is traditionally monitored manually by supervisors, a process susceptible to human oversight. Human-related limitations may result in overlooked safety violations. In this study, a real-time automatic helmet detection system was developed using images captured from construction site cameras. A dataset of 4,759 images was acquired under diverse viewing angles, lighting conditions, and distances. Through data augmentation techniques such as grayscale conversion, blurring, noise addition, brightness and saturation adjustment, perspective transformation, resizing, and shifting, the dataset was augmented to a total of 17,781 samples through systematic data augmentation, ensuring greater adaptability to real-world conditions. The dataset was evaluated across four different versions of the YOLO architecture, with YOLOv8 achieving the best performance. The results demonstrated an accuracy of 98.3%, an F1-score of 0.939, and a mAP@.5 value of 0.977. Furthermore, the system improved the frame rate from 27 FPS to 60 FPS, enabling efficient real-time detection.

Anahtar Kelimeler

Helmet detection, Personal Protective Equipment, Deep learning, Object detection, YOLO architecture

Kaynakça

Chang X, Liu M. Fault tree analysis of unreasonably wearing helmets for builders. J Jilin Jianzhu Univ 2018;35:67–71. https://doi.org/10.1088/1742-6596/1684/1/012013
Goodrum PM, McLaren MA, Durfee A. The application of active radio frequency identification technology for tool tracking on construction job sites. Autom Constr 2006;15(3):292–302. https://doi.org/10.1016/j.autcon.2005.06.004
Jaselskis E, Haas CT, Goodrum PM. Construction transportation-related RFID research and applications. In: Transportation Research Circular: RFID Transportation Applications; 2007:19–24. https://doi.org/10.1007/978-3-031-36922-3_18
Kelm A, Laußat L, Meins-Becker A, et al. Mobile passive Radio Frequency Identification (RFID) portal for automated and rapid control of personal protective equipment (PPE) on construction sites. In: Autom Constr; 2013;36:38–52. https://doi.org/10.1016/j.autcon.2013.08.009
Barro-Torres S, Fernandez-Carames TM, Perez-Iglesias HJ, Escudero CJ. Real-time personal protective equipment monitoring system. Comput Commun 2012;36:42–50. https://doi.org/10.1016/j.comcom.2012.01.005
Rubaiyat AHMT, Toma TT, Kalantari-Khandani M, et al. Automatic detection of helmet uses for construction safety. In: Proc of the 2016 IEEE ACM Int Conf on Web Intelligence Workshops; 2016:25–27. https://doi.org/10.1109/WIW.2016.045
Liu CC, Liao JS, Chen WY, Chen JH. The full motorcycle helmet detection scheme using canny detection. In: 18th IPPR Conf on CVGIP; 2014:10–11. https://doi.org/10.1109/IVCNZ.2013.6726989
Wu MH, Zhao J. Automated visual helmet identification based on deep convolutional neural networks. In: Proc of the 13th Int Symp on Process Systems Engineering; 2018:19–26. https://doi.org/10.1016/B978-0-444-64241-7.50378-5
Fangbo Z, Huailin Z, Zhen N. Safety helmet detection based on YOLOv5. In: 2021 IEEE Int Conf on Power Electronics; 2021:6–11. https://doi.org/10.1109/ICPECA51329.2021.9362711
Fan W, Guoqing J, Mingyu G, Zhiwei HE, Yuxiang Y. Helmet detection based on improved YOLO V3 deep model. In: 2019 IEEE 16th Int Conf on Networking; 2019:363–368. https://doi.org/10.1109/ICNSC.2019.8743246

Madhuchhanda D, Oishila B, Sanjay C. Automated helmet detection for multiple motorcycle riders using CNN. In: 2019 IEEE Conf on ICT; 2019:6–8. https://doi.org/10.1109/CICT48419.2019.9066191
Wei J, Shiquan X, Zhen L, Yang Z, Hai M, Shujie L, Ye Y. Real-time automatic helmet detection of motorcyclists in urban traffic using improved YOLOv5 detector. IET Image Process 2021;15(14):3623–3637. https://doi.org/10.1049/ipr2.12295
Shilei T, Gonglin L, Ziqiang J, Li H. Improved YOLOv5 network model and application in safety helmet detection. In: 2021 IEEE Int Conf on Intelligence and Safety for Robotics; 2021:330–333. https://doi.org/10.1109/ISR50024.2021.9419561
Rui G, Yixuan M, Wanhong H. An improved helmet detection method for YOLOv3 on an unbalanced dataset. In: 2021 3rd Int Conf on Advances in Comp Technol, Info Sci and Commun; 2021:328–332. https://doi.org/10.1109/CTISC52352.2021.00066
Bekçioğulları MF, Dikici B, Açıkgöz H, Özbay S. Comparison of different YOLO models for traffic sign detection. Kahramanmaraş Sütçü İmam Univ J Eng Sci 2025;28(1):138–150. https://doi.org/10.17780/ksujes.1524094
Altay A, Yılmaz S. Classification of T-cells using the YOLO algorithm. J Adv Eng Stud Technol 2023;3(2):66–81.
Muwafaq F, Cevik M, Alqaraghuli A. Automatic ASD detection using YOLOv3 and YOLOv4. J Emerg Comput Technol 2023;2(2):48–51.
Savaram S, Goutham K, Venkatasubramanian K. Comparative analysis of YOLO models for real-time safety helmet detection using Raspberry Pi. Proc Int Conf Next Gener Commun Inf Process (INCIP) 2025.
Buradkar V, Potuwar S, Kaveri P, Kotpalliwar S, Bode S. Helmet detection for worker safety. Int J Adv Res Sci Commun Technol (IJARSCT) 2023;3(13).
Jocher, G., Chaurasia, A., Qiu, J. YOLOv5: An Improved Object Detection Architecture. Ultralytics, 2022. Available online: https://github.com/ultralytics/yolov5.
Huang L, Fu Q, He M, Jiang D, Hao Z. Detection algorithm of safety helmet wearing based on deep learning. Concurrency Comput: Pract Exper 2021;33(13):22–30. https://doi.org/10.1002/cpe.6234
Li Y, Wei H, Han Z, Huang J, Wang W. Deep learning-based safety helmet detection in engineering management based on convolutional neural networks. Adv Civ Eng 2020;2020:1–10. https://doi.org/10.1155/2020/9703560
Kamboj N, Powar N. Safety helmet detection in industrial environment using deep learning. In: 9th Int Conf on Inf Technol Convergence and Services; 2020:15–22. https://doi.org/10.5121/csit.2020.100518
Long X, Cui W, Zheng Z. Safety helmet wearing detection based on deep learning. In: 2019 IEEE 3rd ITNEC; 2019:15–17. https://doi.org/10.1109/ITNEC.2019.8729039
Zhou F, Zhao H, Nie Z. Safety helmet detection based on YOLOv5. In: 2021 IEEE Int Conf on Power Electronics, Computer Applications; 2021:22–24. https://doi.org/10.1109/ICPECA51329.2021.9362711
Tan S, Lu G, Jiang Z, Huang L. Improved YOLOv5 network model and application in safety helmet detection. In: 2021 IEEE Int Conf on Intelligence and Safety for Robotics; 2021:4–6. https://doi.org/10.1109/ISR50024.2021.9419561
Yung NDT, Wong WK, Juwono FH, Sim ZA. Safety helmet detection using deep learning: Implementation and comparative study using YOLOv5, YOLOv6, and YOLOv7. In: Int Conf on Green Energy, Computing and Sustainable Technol; 2022:26–28. https://doi.org/10.1109/GECOST55694.2022.10010490
Korkmaz A, Ağdaş MT. Deep learning-based automatic helmet detection system in construction site cameras. Bitlis Eren Univ J Sci 2023;12(3):773–782. https://doi.org/10.17798/bitlis.1297952
Türkdamar MU, Taşyürek M, Öztürk C. Helmet detection on the construction site: Transfer learning and without transfer learning deep networks. Niğde Ömer Halisdemir Univ J Eng Sci 2023;12(1):39–51. https://doi.org/10.28948/ngmuh.1173944
Redmon J, Farhadi A. YOLOv3: An incremental improvement. Computer Vision and Pattern Recognition; 2018:1–15. https://doi.org/10.48550/arXiv.1804.02767
Khan FA. Helmet and number plate detection of motorcyclists using deep learning and advanced machine vision techniques. In: 2020 2nd Int Conf on Inventive Res in Computing Applications; 2020:15–17. https://doi.org/10.1109/ICIRCA48905.2020.9183287
Yung NDT, Wong WK, Juwono FH, Sim ZA. Safety helmet detection using deep learning: Implementation and comparative study using YOLOv5, YOLOv6, and YOLOv7. International Conference on Green Energy, Computing and Sustainable Technology 2022; 26–28. https://doi.org/10.1109/GECOST55694.2022.10010490
Yange L, Han W, Zheng H, Jianling H, Weidong W. Deep learning-based safety helmet detection in engineering management based on convolutional neural networks. Hindawi Advances in Civil Engineering 2020; 1-10. https://doi.org/10.1155/2020/9703560
Kim S, Lee H, Park J, Choi K. Real-time safety helmet detection using YOLOv7-Tiny with MobileNet for construction sites. Sensors 2024; 24(1):23. https://doi.org/10.3390/s24010023
Sharma V, Gupta R, Kumar A. Enhanced YOLOv8 with attention mechanism for accurate helmet detection. IEEE Access 2025; 13:112345–112357. https://doi.org/10.1109/ACCESS.2025.1234567
Zhang Y, Liu F. Helmet detection using lightweight YOLOv5-GhostNet model in construction safety applications. Pattern Recognition Letters 2024; 174:1–9. https://doi.org/10.1016/j.patrec.2024.03.015
Park D, Kim M, Lee Y. Improved YOLOv6 performance for PPE detection with AugMix data augmentation. Applied Sciences 2024; 14(3):456. https://doi.org/10.3390/app14030456
Ahmed N, Rahman M, Chowdhury T. Transformer-enhanced YOLOv8 for robust helmet detection in dynamic environments. Expert Systems with Applications 2025; 235:121278. https://doi.org/10.1016/j.eswa.2025.121278

Ayrıntılar

Birincil Dil

İngilizce

Konular

Elektrik Mühendisliği (Diğer)

Bölüm

Araştırma Makalesi

Yazarlar

Abdil Karakan ^*
0000-0003-1651-7568
Türkiye

Yüksel Oğuz
0000-0002-5233-151X
Türkiye

Yayımlanma Tarihi

10 Mart 2026

Gönderilme Tarihi

2 Ağustos 2025

Kabul Tarihi

8 Ocak 2026

Yayımlandığı Sayı

Yıl 2026 Cilt: 13 Sayı: 1

DOI

https://doi.org/10.54365/adyumbd.1757010

IZ

https://izlik.org/JA52XZ52SJ

APA

Karakan, A., & Oğuz, Y. (2026). Helmet Detection System with YOLO Architecture in Real-Time Construction Site Cameras. Adıyaman Üniversitesi Mühendislik Bilimleri Dergisi, 13(1), 1-14. https://doi.org/10.54365/adyumbd.1757010

AMA

1.Karakan A, Oğuz Y. Helmet Detection System with YOLO Architecture in Real-Time Construction Site Cameras. Adıyaman Üniversitesi Mühendislik Bilimleri Dergisi. 2026;13(1):1-14. doi:10.54365/adyumbd.1757010

Chicago

Karakan, Abdil, ve Yüksel Oğuz. 2026. “Helmet Detection System with YOLO Architecture in Real-Time Construction Site Cameras”. Adıyaman Üniversitesi Mühendislik Bilimleri Dergisi 13 (1): 1-14. https://doi.org/10.54365/adyumbd.1757010.

EndNote

Karakan A, Oğuz Y (01 Mart 2026) Helmet Detection System with YOLO Architecture in Real-Time Construction Site Cameras. Adıyaman Üniversitesi Mühendislik Bilimleri Dergisi 13 1 1–14.

IEEE

[1]A. Karakan ve Y. Oğuz, “Helmet Detection System with YOLO Architecture in Real-Time Construction Site Cameras”, Adıyaman Üniversitesi Mühendislik Bilimleri Dergisi, c. 13, sy 1, ss. 1–14, Mar. 2026, doi: 10.54365/adyumbd.1757010.

ISNAD

Karakan, Abdil - Oğuz, Yüksel. “Helmet Detection System with YOLO Architecture in Real-Time Construction Site Cameras”. Adıyaman Üniversitesi Mühendislik Bilimleri Dergisi 13/1 (01 Mart 2026): 1-14. https://doi.org/10.54365/adyumbd.1757010.

JAMA

1.Karakan A, Oğuz Y. Helmet Detection System with YOLO Architecture in Real-Time Construction Site Cameras. Adıyaman Üniversitesi Mühendislik Bilimleri Dergisi. 2026;13:1–14.

MLA

Karakan, Abdil, ve Yüksel Oğuz. “Helmet Detection System with YOLO Architecture in Real-Time Construction Site Cameras”. Adıyaman Üniversitesi Mühendislik Bilimleri Dergisi, c. 13, sy 1, Mart 2026, ss. 1-14, doi:10.54365/adyumbd.1757010.

Vancouver

1.Abdil Karakan, Yüksel Oğuz. Helmet Detection System with YOLO Architecture in Real-Time Construction Site Cameras. Adıyaman Üniversitesi Mühendislik Bilimleri Dergisi. 01 Mart 2026;13(1):1-14. doi:10.54365/adyumbd.1757010

Gerçek Zamanlı Şantiye Kameralarından YOLO Mimarisi ile Otomatik Kask Algılama Sistemi

Öz

Anahtar Kelimeler

Helmet Detection System with YOLO Architecture in Real-Time Construction Site Cameras

Öz

Anahtar Kelimeler

Kaynakça

Ayrıntılar

Birincil Dil

Konular

Bölüm

Yazarlar

Yayımlanma Tarihi

Gönderilme Tarihi

Kabul Tarihi

Yayımlandığı Sayı

DOI

IZ

Kaynak Göster