Sistem Deteksi Pelanggaran Penggunaan Masker Pada Pekerja Area Pasting Untuk Pencegahan Paparan Timbal Menggunakan YOLOv8

Edianandra, Mochamad Eren (2026) Sistem Deteksi Pelanggaran Penggunaan Masker Pada Pekerja Area Pasting Untuk Pencegahan Paparan Timbal Menggunakan YOLOv8. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040221118-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (10MB) | Request a copy

Abstract

Paparan timbal pada area pasting PT GS Battery Semarang merupakan salah satu risiko keselamatan dan kesehatan kerja (K3) yang tinggi akibat masih ditemukannya ketidakpatuhan terhadap penggunaan masker. Pengawasan yang masih dilakukan secara manual melalui CCTV menyebabkan pelanggaran sulit terdeteksi secara konsisten dan tidak terdokumentasi dengan baik. Penelitian ini bertujuan mengembangkan sistem deteksi otomatis berbasis YOLOv8 untuk memonitor kepatuhan penggunaan masker secara real time dan melakukan pencatatan pelanggaran. Sistem diintegrasikan dengan Region of Interest (ROI), algoritma pelacakan ByteTrack untuk memberikan identitas unik pada setiap pekerja, serta Programmable Logic Controller (PLC) untuk mengaktifkan warning light secara otomatis ketika pelanggaran terdeteksi. Hasil pengujian menunjukkan bahwa konfigurasi model baseline dengan pemberian image enhancement histogram equalization pada data latih memberikan peningkatan performa pendeteksian dan generalisasi terhadap data baru pada dataset uji area pasting, dengan capaian Recall sebesar 87,37%, F1-Score sebesar 84,64%, dan Accuracy sebesar 86,24%. Pengujian sistem pada saat proses produksi yang berlangsung selama lima hari, yaitu pada periode 6–10 Juli, mencatat 1.494 kejadian pelanggaran dengan rata-rata akurasi deteksi sebesar 93,46% dibandingkan dengan hasil observasi manual. Meskipun terjadi penurunan frame rate akibat beban komputasi, sistem tetap mampu bekerja secara real time dan menghasilkan dokumentasi pelanggaran secara otomatis.
===================================================================================================================================
Lead exposure in the pasting area of PT GS Battery Semarang poses a significant occupational safety and health (OSH) hazard due to persistent non-compliance with mandatory respirator mask usage. The current surveillance process, which relies on manual monitoring through Closed-Circuit Television (CCTV), often results in inconsistent violation detection and inadequate documentation. This study proposes an automated respirator mask compliance monitoring system based on the YOLOv8 object detection model to enable real-time detection and automatic recording of violations. The proposed system integrates a Region of Interest (ROI), the ByteTrack multi-object tracking algorithm to assign a unique identity to each worker, and a Programmable Logic Controller (PLC) to automatically activate a warning light whenever a violation is detected. Experimental results demonstrate that the baseline YOLOv8 model enhanced with histogram equalization achieved the best detection performance and superior generalization capability on the pasting area test dataset, yielding a Recall of 87.37%, an F1-Score of 84.64%, and an Accuracy of 86.24%. Furthermore, during a five-day production trial conducted from July 6 to July 10, the proposed system recorded 1,494 violation events and achieved an average detection accuracy of 93.46% compared with manual observations. Although the additional computational workload reduced the frame rate, the system maintained real-time performance while automatically documenting detected violations, demonstrating its potential to improve respirator mask compliance monitoring and reduce workers' exposure to lead in industrial environments.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Deteksi Penggunaan Masker, Keselamatan dan Kesehatan Kerja (K3), image enhancement, YOLOv8, Image Enhancement, Mask Compliance Detection, Occupational Health and Safety (OHS), YOLOv8.
Subjects:	Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Mochamad Eren Edianandra
Date Deposited:	10 Aug 2026 02:42
Last Modified:	10 Aug 2026 02:42
URI:	http://repository.its.ac.id/id/eprint/144156

Actions (login required)

View Item