Sistem Deteksi Ketepatan Penggunanan APD Sebagai Autentikasi Akses Menggunakan YOLOv11n_OBB dan YOLO-Pose

Fauzi, Ahmad (2026) Sistem Deteksi Ketepatan Penggunanan APD Sebagai Autentikasi Akses Menggunakan YOLOv11n_OBB dan YOLO-Pose. Other thesis, Institut Teknologi Sepuluh Nopember.

Text 2040221051-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy

Abstract

Keselamatan dan Kesehatan Kerja (K3) merupakan aspek yang sangat krusial dan fundamental dalam melindungi setiap pekerja dari berbagai potensi bahaya di lingkungan industri modern. Penggunaan Alat Pelindung Diri (APD) dasar seperti helm keselamatan, masker, dan sepatu khusus menjadi upaya mitigasi yang esensial. Akan tetapi, tingkat kepatuhan para pekerja di lapangan sering kali menurun akibat pemakaian yang tidak sesuai dengan standar operasional prosedur, seperti posisi helm yang terpasang miring atau penggunaan masker yang tidak menutupi area saluran pernapasan secara sempurna. Untuk mengatasi permasalahan tersebut, penelitian ini berhasil mengembangkan sebuah sistem deteksi dan validasi otomatis penggunaan APD berbasis teknologi computer vision dengan cara mengintegrasikan dua arsitektur deep learning mutakhir. Model YOLOv11-OBB (Oriented Bounding Box) diimplementasikan secara khusus untuk mendeteksi objek APD beserta kemiringannya, sementara model YOLO-Pose digunakan sebagai pengganti metode mesh konvensional guna mengekstraksi keypoints anatomi penting tubuh manusia, seperti area hidung dan pergelangan kaki, untuk mengevaluasi ketepatan posisi pemasangan APD secara presisi. Berdasarkan hasil pengujian eksperimental yang telah dilakukan, evaluasi menunjukkan bahwa arsitektur yang diusulkan mampu mencapai tingkat akurasi sistem deteksi yang sangat baik, yakni mencapai 92% terhadap variasi jarak pengamatan sejauh 2 hingga 2,5 meter. Selain itu, berdasarkan hasil pengujian performa yang menggunakan perangkat keras lokal berbasis CPU Intel Core i3, sistem ini mampu mencapai kecepatan pemrosesan hingga 13 FPS secara stabil. Melalui implementasi sistem cerdas ini, diharapkan dapat segera mewujudkan sistem deteksi APD yang terotomatisasi secara penuh, efisien, serta sangat responsif guna mendukung terciptanya lingkungan kerja industri yang aman dan terlindungi bagi seluruh pekerja.
====================================================================================================================================
Occupational Health and Safety (OHS) is a crucial and fundamental aspect of protecting every worker from various potential hazards in modern industrial environments. The use of basic Personal Protective Equipment (PPE) such as safety helmets, masks, and specialized boots serves as an essential mitigation effort. However, compliance levels among workers in the field often decline due to usage that does not comply with standard operating procedures, such as helmets installed at an angle or masks failing to properly cover the respiratory area. To address these issues, this research successfully develops an automated PPE detection and validation system based on computer vision technology by integrating two advanced deep learning architectures. The YOLOv11-OBB (Oriented Bounding Box) model is specifically implemented to accurately detect PPE objects along with their tilt angles, while the YOLO-Pose model is utilized as a replacement for conventional mesh methods to extract important human anatomical keypoints, such as the nose and ankle areas, to precisely evaluate the correctness of PPE positioning. Based on experimental testing results, the evaluation shows that the proposed architecture achieves an impressive detection system accuracy rate of 92% against observation distance variations ranging from 2 to 2.5 meters. Furthermore, based on performance testing using local hardware powered by an Intel Core i3 CPU, the system achieves a stable processing speed of up to 13 FPS. Through the implementation of this intelligent system, it is expected to establish a fully automated, efficient, and highly responsive PPE detection system to support the creation of a safe and protected industrial work environment for all employees.

Item Type:	Thesis (Other)
Uncontrolled Keywords:	Keselamatan dan Kesehatan Kerja, Alat Pelindung Diri (APD), YOLOv11-OBB, YOLO-Pose, Computer Vision, Antarmuka Pengguna, Personal Protective Equipment (PPE), Computer Vision, YOLOv11-OBB, YOLO-Pose, Automated Validation.
Subjects:	T Technology > T Technology (General) > T55 Industrial Safety T Technology > TA Engineering (General). Civil engineering (General) > TA1573 Detectors. Sensors T Technology > TA Engineering (General). Civil engineering (General) > TA1637 Image processing--Digital techniques. Image analysis--Data processing. T Technology > TA Engineering (General). Civil engineering (General) > TA1650 Face recognition. Optical pattern recognition.
Divisions:	Faculty of Vocational > 36304-Automation Electronic Engineering
Depositing User:	Ahmad Fauzi
Date Deposited:	14 Aug 2026 02:43
Last Modified:	14 Aug 2026 02:43
URI:	http://repository.its.ac.id/id/eprint/144340

Actions (login required)

View Item