Rancang Bangun Sistem Pendeteksi Alat Pelindung Diri (Compliant) Berbasis Kamera Dan Metode You Only Look Once (YOLO) Di PT Pertamina EP Sukowati Field

Azhari, M Dwi Aswangga Azhari (2026) Rancang Bangun Sistem Pendeteksi Alat Pelindung Diri (Compliant) Berbasis Kamera Dan Metode You Only Look Once (YOLO) Di PT Pertamina EP Sukowati Field. Project Report. [s.n]. (Unpublished)

Text 2042221103-Project_Report.pdf - Accepted Version Restricted to Repository staff only Download (8MB) | Request a copy

Abstract

Industri minyak dan gas memiliki tingkat risiko kerja yang tinggi sehingga penerapan Keselamatan dan Kesehatan Kerja (K3), khususnya penggunaan Alat Pelindung Diri (APD), menjadi aspek yang sangat penting. Pada PT Pertamina EP Sukowati Field, pengawasan penggunaan APD masih dilakukan secara manual sehingga kurang efektif untuk pemantauan secara real-time dan berpotensi menimbulkan kelalaian. Penelitian ini bertujuan untuk merancang dan membangun sistem pendeteksi APD berbasis kamera menggunakan metode You Only Look Once (YOLO). Sistem yang dikembangkan mampu mendeteksi lima jenis APD, yaitu safety helmet, coverall, safety glasses, safety shoes, dan gloves, serta mengklasifikasikan kondisi pekerja ke dalam kategori compliant dan non-compliant. Tahapan penelitian meliputi pengumpulan dataset, pelabelan data menggunakan Roboflow, pelatihan model menggunakan Google Colab, serta pengembangan perangkat lunak berbasis Python dengan dukungan OpenCV dan YOLO. Sistem juga diintegrasikan dengan kamera CCTV, Arduino Uno, lampu rotary, dan speaker sebagai aktuator peringatan. Hasil penelitian menunjukkan model memperoleh nilai F1-score optimum sebesar 0,76 pada confidence threshold 0,487 dan nilai mAP@0.5 sebesar 75,3%. Pada pengujian, sistem mampu mendeteksi APD secara real-time dengan detection rate pada rentang 72%.
================================================================================================================================
The oil and gas industry involves high occupational risks, making the implementation of Occupational Safety and Health (OSH)—particularly the use of Personal Protective Equipment (PPE)—a critical aspect. At PT Pertamina EP Sukowati Field, monitoring of PPE use is still conducted manually, which limits its effectiveness for real-time monitoring and may lead to oversights. This study aims to design and develop a camera-based PPE detection system using the You Only Look Once (YOLO) method. The developed system is capable of detecting five types of PPE—safety helmets, protective suits, safety goggles, safety shoes, and gloves—and classifying workers’ compliance status into compliant and non-compliant categories. The research stages included dataset collection, data labeling using Roboflow, model training using Google Colab, and the development of Python-based software supported by OpenCV and YOLO. The system was also integrated with a CCTV camera, an Arduino Uno, a rotating light, and a speaker as warning actuators. The results show that the model achieved an optimal F1-score of 0.76 at a confidence threshold of 0.487 and a mAP@0.5 of 75.3%. During testing, the system was able to detect PPE in real-time with a detection rate in the range of 72%.

Item Type:	Monograph (Project Report)
Uncontrolled Keywords:	Alat Pelindung Diri, K3, Object Detection, YOLO ============================================================ Personal Protective Equipment, Occupational Safety and Health, Object Detection, YOLO
Subjects:	T Technology > T Technology (General) > T55 Industrial Safety T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques. Image analysis--Data processing.
Divisions:	Faculty of Vocational > Instrumentation Engineering
Depositing User:	M Dwi Aswangga Azhari
Date Deposited:	29 Jul 2026 08:17
Last Modified:	29 Jul 2026 08:17
URI:	http://repository.its.ac.id/id/eprint/139050

Actions (login required)

View Item