Implementasi Sistem Deteksi Objek Penggunaan APD Menggunakan Algoritma YOLO Berbasis Raspberry PI Untuk PT Pertamina Patra Niaga

Mukminin, Bintang Amirul (2026) Implementasi Sistem Deteksi Objek Penggunaan APD Menggunakan Algoritma YOLO Berbasis Raspberry PI Untuk PT Pertamina Patra Niaga. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2043221017-Undergraduate_Thesis.pdf Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Keselamatan dan Kesehatan Kerja (K3) merupakan aspek penting dalam dunia industri untuk melindungi pekerja dari potensi bahaya di lingkungan kerja. Salah satu elemen utama penerapan K3 adalah kepatuhan terhadap penggunaan Alat Pelindung Diri (APD), namun proses pemeriksaan manual di lapangan masih memiliki keterbatasan efisiensi dan akurasi. Penelitian ini bertujuan mengembangkan sistem deteksi otomatis penggunaan APD berbasis computer vision menggunakan algoritma YOLOv8 yang diimplementasikan pada perangkat Raspberry Pi 5 dengan dukungan akselerator Hailo-8 di lingkungan kerja PT Pertamina Patra Niaga Integrated Terminal Surabaya. Data penelitian terdiri dari 4.257 citra pekerja industri dengan lima kelas objek yaitu Helmet, No-Helmet, No-Safety-Boot, Person, dan Safety-Boot. Seluruh data melalui proses preprocessing dan augmentasi untuk meningkatkan kemampuan generalisasi model. Hasil pelatihan model YOLOv8s menunjukkan performa yang baik dengan nilai precision sebesar 86,8%, recall 81,8%, dan mAP50 87,4%. Pengujian terhadap 412 citra menghasilkan precision 89,3%, recall 82,4%, dan mAP50 88,7% dengan F1-score sebesar 0,86 yang menandakan keseimbangan antara precision dan recall. Implementasi sistem secara real-time pada Raspberry Pi 5 menunjukkan performa stabil dengan kecepatan 27–29 FPS serta akurasi 91% pada pagi hari dan 89% pada sore hari. Sistem juga dilengkapi fitur voice response dan notifikasi otomatis melalui Telegram Bot. Hasil penelitian ini membuktikan bahwa integrasi YOLOv8s pada Raspberry Pi 5 dan Hailo-8 mampu mendeteksi penggunaan APD secara otomatis dan efisien untuk meningkatkan efektivitas pengawasan K3 di PT Pertamina Patra Niaga.
====================================================================================================================================
Health, Safety, and Environment (HSE) is an essential aspect of the industrial sector to protect workers from potential hazards in the workplace. One of the key elements in implementing HSE is compliance with the use of Personal Protective Equipment (PPE), although manual inspection processes in the field still face limitations in efficiency and accuracy. This study aims to develop an automatic PPE detection system based on computer vision using the YOLOv8 algorithm, implemented on a Raspberry Pi 5 device with Hailo-8 accelerator support in the working environment of PT Pertamina Patra Niaga Integrated Terminal Surabaya. The dataset consists of 4,257 industrial worker images across five object classes with Helmet, No-Helmet, No-Safety-Boot, Person, and Safety-Boot. All data underwent preprocessing and augmentation to enhance model generalization capability. The training results of the YOLOv8s model demonstrated good performance, achieving a precision of 86.8%, recall of 81.8%, and mAP50 of 87.4%. Testing on 412 images yielded a precision of 89.3%, recall of 82.4%, and mAP50 of 88.7%, with an F1-score of 0.86, indicating a strong balance between precision and recall. The real-time implementation on Raspberry Pi 5 showed stable performance at 27–29 FPS with an accuracy of 91% in the morning and 89% in the afternoon. The system is also equipped with voice response and automatic notifications via Telegram Bot. The results prove that integrating YOLOv8s with Raspberry Pi 5 and Hailo-8 can efficiently and automatically detect PPE usage, enhancing the effectiveness of HSE supervision at PT Pertamina Patra Niaga.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	APD, Computer Vision, Hailo-8, Raspberry Pi, Roboflow, YOLOv8, PPE, Computer Vision, Hailo-8, Raspberry Pi, Roboflow, YOLOv8.
Subjects:	Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) Q Science > QA Mathematics > QA76.9 Computer algorithms. Virtual Reality. Computer simulation. T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques. Image analysis--Data processing.
Divisions:	Faculty of Vocational > 49501-Business Statistics
Depositing User:	Bintang Amirul Mukminin
Date Deposited:	10 Jun 2026 00:41
Last Modified:	10 Jun 2026 00:41
URI:	http://repository.its.ac.id/id/eprint/133536

Actions (login required)

View Item