Performance Analysis of Various CNN Models for Detecting Personal Protective Equipment on Construction Workers
Keywords:
Deteksi APD, CNN, YOLO, Faster R-CNN, PerfomaAbstract
Industri konstruksi terus menghadapi tantangan keselamatan yang signifikan, dengan tingginya angka kecelakaan kerja yang sering kali terkait dengan penggunaan alat pelindung diri (APD) yang tidak memadai. Secara global, sektor konstruksi menyumbang 31,9% dari kecelakaan kerja. Di Indonesia, tercatat 370.747 kecelakaan kerja pada tahun 2023, termasuk 2.965 kasus di sektor konstruksi, yang menunjukkan adanya celah besar dalam protokol keselamatan. Metode pemantauan APD konvensional tidak lagi memadai di lingkungan kerja yang dinamis, sehingga dibutuhkan sistem deteksi cerdas. Penelitian ini bertujuan untuk: (1) mengembangkan model deteksi APD menggunakan jaringan saraf konvolusional (CNN), dan (2) mengevaluasi berbagai model CNN untuk mengidentifikasi yang paling efektif. Lima model berbasis CNN, YOLOv9, YOLOv10, YOLOv11, Faster R-CNN-MobileNet, dan Faster R-CNN-EfficientNet dikembangkan dan diuji menggunakan 3.000 gambar beranotasi yang mencakup item APD seperti helm, rompi, masker, sarung tangan, dan sepatu. Dengan kerangka kerja R&D berbasis ADDIE, model dievaluasi menggunakan metrik presisi, recall, F1-score, mAP, waktu inferensi, dan FPS. Analisis komparatif menunjukkan YOLOv11 memiliki performa terbaik, dengan mAP 93,2% dan inferensi waktu nyata pada 11 FPS. Hasil ini menunjukkan efektivitas YOLOv11 untuk lokasi konstruksi dengan sumber daya terbatas, memungkinkan deteksi dini pelanggaran APD dan meningkatkan kepatuhan terhadap peraturan. Inovasi ini mendukung upaya pencegahan kecelakaan melalui identifikasi ketidakpatuhan lebih awal.
Downloads
References
Akram, R., Thaheem, M. J., Khan, S., Nasir, A. R., & Maqsoom, A. (2022). Exploring the Role of BIM in Construction Safety in Developing Countries: Toward Automated Hazard Analysis. Sustainability (Switzerland), 14(19). https://doi.org/10.3390/su141912905
Amangeldy, N., Barlybaev, A., Gazizova, N., & Kurmetbek, B. (2024). Evaluation of YOLOv8 and YOLOv10 Models in PPE Recognition Tasks. Journal of Electrical Systems, 20(10s), 8141–8148.
Arfa, A. N., Julaeha, S., Awaludin, A. A. R., Priambodo, C. G., & Satria, H. (2023). PENGEMBANGAN INSTRUMEN PEMBELAJARAN MENGGUNAKAN VIDEO HOLOGRAM 3D DALAM MASA HYBIRD. Jurnal Indonesia : Manajemen Informatika Dan Komunikasi, 4(1), 343–348. https://doi.org/10.35870/jimik.v4i1.198
Barlybayev, A., Amangeldy, N., Kurmetbek, B., Krak, I., Razakhova, B., Tursynova, N., & Turebayeva, R. (2024). Personal protective equipment detection using YOLOv8 architecture on object detection benchmark datasets: a comparative study. Cogent Engineering, 11(1). https://doi.org/10.1080/23311916.2024.2333209
BPJS Ketenagakerjaan. (2024). Kecelakaan Kerja Makin Marak dalam Lima Tahun Terakhir. BPJS Ketenagakerjaan. https://www.bpjsketenagakerjaan.go.id/berita/28681/Kecelakaan-Kerja-Makin-Marak-dalam-Lima-Tahun-Terakhir
Feng, R., Miao, Y., & Zheng, J. (2024). A YOLO-based intelligent detection algorithm for risk assessment of construction sites. Journal of Intelligent Construction. https://doi.org/10.26599/JIC.2024.9180037
Guney, E., Altin, H., Esra Asci, A., Bayilmis, O. U., & Bayilmis, C. (2024). YOLO-Based Personal Protective Equipment Monitoring System for Workplace Safety. JITSI : Jurnal Ilmiah Teknologi Sistem Informasi, 5(2), 77–85. https://doi.org/10.62527/jitsi.5.2.238
Harahap, F. G. S. (2023). Factors Related To Compliance With The Use Of Personal Protective Equipment (PPE) Against Employees At Pt.Mas Batang Toru Mine 2023. International Journal of Health Engineering and Technology, 1(6). https://doi.org/10.55227/ijhet.v1i6.113
Ke, X., Chen, W., & Guo, W. (2022). 100+ FPS detector of personal protective equipment for worker safety: A deep learning approach for green edge computing. Peer-to-Peer Networking and Applications, 15(2), 950–972. https://doi.org/10.1007/s12083-021-01258-4
Kementerian Ketenagakerjaan Republik Indonesia. (2024, July 11). Kasus Kecelakaan Kerja, Maret Tahun 2024. Satu Data Kementerian Ketenagakerjaan. https://satudata.kemnaker.go.id/data/kumpulan-data/1875
LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521(7553), 436–444. https://doi.org/10.1038/nature14539
Liu, Y., & Jiang, W. (2021). Detection of wearing safety helmet for workers based on YOLOv4. 2021 International Conference on Computer Engineering and Artificial Intelligence (ICCEAI), 83–87. https://doi.org/10.1109/ICCEAI52939.2021.00016
Molenda, M. (2003). In search of the elusive ADDIE model. Performance Improvement, 42(5), 34–36. https://doi.org/10.1002/pfi.4930420508
Mostofi, F., Toğan, V., Ayözen, Y. E., & Tokdemir, O. B. (2022). Construction Safety Risk Model with Construction Accident Network: A Graph Convolutional Network Approach. Sustainability (Switzerland), 14(23). https://doi.org/10.3390/su142315906
Muthaqin, D. I. (2023). Frame Work Occupational Accident-Based Labor Law Enforcement in the Protection of Labor Rights. Proceedings of the 4th Annual Civic Education Conference (ACEC 2022), 146–154. https://doi.org/10.2991/978-2-38476-096-1_17
Protik, A. A., Rafi, A. H., & Siddique, S. (2021). Real-time Personal Protective Equipment (PPE) Detection Using YOLOv4 and TensorFlow. 2021 IEEE Region 10 Symposium (TENSYMP), 1–6. https://doi.org/10.1109/TENSYMP52854.2021.9550808
Ren, D., Sun, T., Yu, C., & Zhou, C. (2021). Research on Safety Helmet Detection for Construction Site. 2021 International Conference on Computer Information Science and Artificial Intelligence (CISAI), 186–189. https://doi.org/10.1109/CISAI54367.2021.00042
Saudi, M. M., Hakim, A., Ahmad, A., Shakir, A., Hanafi, M., Narzullaev, A., & Ifwat, M. (2020). Image Detection Model for Construction Worker Safety Conditions using Faster R-CNN. International Journal of Advanced Computer Science and Applications, 11(6). https://doi.org/10.14569/IJACSA.2020.0110632
Setiawan, E., Witjaksana, B., & Tjendani, H. T. (2023). Analysis of Risk Management in Building Workers of SMAN 5 Brawijaya Building Kediri. International Journal on Advanced Technology, Engineering, and Information System (IJATEIS), 2(4).
Shen, S., & Yang, J. (2024). Better YOLO with Attention-Augmented Network and Enhanced Generalization Performance for Safety Helmet Detection.
Shi, C., Zhu, D., Shen, J., Zheng, Y., & Zhou, C. (2023). GBSG-YOLOv8n: A Model for Enhanced Personal Protective Equipment Detection in Industrial Environments. Electronics, 12(22), 4628. https://doi.org/10.3390/electronics12224628
Sugiyono. (2019). Metode Penelitian Kuantitatif, Kualitatif, dan R&D. Alfabeta.
Downloads
Published
Issue
Section
License

This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.