Yuniardi, Ismail Adrian Putra (2026) Sistem Deteksi Penggunaan Alat Pelindung Diri dan Pengenalan Identitas Pekerja Menggunakan YOLOv11 dan Optical Character Recognition (OCR). Other thesis, Institut Teknologi Sepuluh Nopember.
![[thumbnail of 2040221084-Undergraduate_Thesis.pdf]](http://repository.its.ac.id/style/images/fileicons/text.png) |
Text
2040221084-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only Download (4MB) | Request a copy |
Abstract
Lingkungan operasional Pembangkit Listrik Tenaga Mesin Gas (PLTMG) memiliki karakteristik risiko tinggi yang melibatkan paparan tegangan menengah-tinggi, kebisingan ekstrem, dan bahaya termal, sehingga menuntut penerapan standar Keselamatan dan Kesehatan Kerja (K3) yang ketat. Saat ini, mekanisme pengawasan kepatuhan penggunaan Alat Pelindung Diri (APD) dan verifikasi akses personel di PLTMG PT. Lamong Energi Indonesia masih mengandalkan pemantauan manual oleh petugas keamanan. Metode konvensional ini memiliki kelemahan fundamental berupa inkonsistensi akibat faktor kelelahan manusia (human fatigue) dan keterbatasan atensi visual dalam pengawasan kontinu 24 jam, yang berpotensi meningkatkan risiko kecelakaan kerja dan akses ilegal. Penelitian ini mengusulkan rancang bangun sistem pengawasan cerdas terintegrasi yang memanfaatkan teknologi Computer Vision dan Deep Learning untuk mengotomatisasi proses tersebut. Sistem dikembangkan menggunakan algoritma Convolutional Neural Network (CNN) dengan arsitektur YOLOv11 untuk deteksi kelengkapan APD (helm dan rompi keselamatan) secara real-time, serta Optical Character Recognition (OCR) untuk digitalisasi dan verifikasi identitas melalui kartu identitas (ID Card). Metodologi penelitian mencakup pengumpulan dataset primer di lingkungan simulasi, pelatihan model menggunakan teknik transfer learning untuk optimalisasi akurasi, dan implementasi perangkat keras yang mengintegrasikan kamera, unit pemrosesan, dan sistem aktuator peringatan dini (alarm dan lampu indikator). Hasil pengujian menunjukkan model YOLOv11 mencapai mean Average Precision (mAP@0.5) sebesar 95,7%, dengan akurasi deteksi rompi dan ID Card 100%, serta helm 91%. Pada modul teks, pra-pemrosesan citra dan filter Regular Expression (Regex) menghasilkan Exact Match Rate puncak 87,10% pada jarak operasional ideal 10 cm. Integrasi sistem mampu merespons dengan waktu komputasi rata-rata 1,74 detik, sementara teknik multithreading menjaga laju kamera tetap stabil pada ~30 Frames Per Second (FPS) tanpa hambatan jeda mekanis gerbang. Sistem terbukti 100% berhasil dalam 30 skenario pengujian, menciptakan lapisan keamanan ganda (defense-in-depth) yang responsif untuk meminimalkan human error dalam penerapan K3 di pembangkit listrik
==================================================================================================================================
Gas Engine Power Plants (PLTMG) are high-risk industrial environments characterized by exposure to medium-to-high voltage, extreme noise, and thermal hazards, necessitating stringent adherence to Occupational Health and Safety (OHS) protocols. Currently, compliance monitoring for Personal Protective Equipment (PPE) usage and personnel access verification at PLTMG PT. Lamong Energi Indonesia relies heavily on manual inspection by security personnel. This conventional method suffers from fundamental limitations, including inconsistency due to human fatigue and restricted visual attention spans during continuous 24-hour surveillance, thereby increasing the probability of workplace accidents and unauthorized access. This study proposes the design and development of an integrated intelligent surveillance system utilizing Computer Vision and Deep Learning technologies to automate these processes. The system employs a Convolutional Neural Network (CNN) based on the YOLOv11 architecture for real-time detection of PPE completeness (safety helmets and vests), alongside Optical Character Recognition (OCR) for identity digitization and verification via ID Cards. The research methodology encompasses primary dataset collection in a simulated environment, model training utilizing transfer learning techniques to optimize accuracy, and hardware implementation integrating a camera, a processing unit, and early warning actuator systems (alarms and indicator lights). Test results demonstrate that the YOLOv11 model achieves a mean Average Precision (mAP@0.5) of 95.7%, with a 100% accuracy rate for safety vest and ID Card detection, and 91% for safety helmets. For the text recognition module, the application of image preprocessing and a Regular Expression (Regex) filter yields a peak Exact Match Rate of 87.10% at an ideal operational distance of 10 cm. The integrated system responds with an average computational time of 1.74 seconds, while the implementation of multithreading techniques maintains a stable camera frame rate at approximately 30 Frames Per Second (FPS) without being hindered by the gate's mechanical delay. The system proved 100% successful across 30 testing scenarios, establishing a responsive defense-in-depth mechanism to minimize human error in the enforcement of OHS protocols within the power plant.
| Item Type: | Thesis (Other) |
|---|---|
| Uncontrolled Keywords: | Alat Pelindung Diri (APD), Convolutional Neural Network (CNN), Deep Learning, Keselamatan dan Kesehatan Kerja (K3), Optical Character Recognition (OCR), PLTMG, YOLOv11, Convolutional Neural Network (CNN), Deep Learning, Gas Engine Power Plant (PLTMG), Occupational Health and Safety (K3), Optical Character Recognition (OCR), Personal Protective Equipment (PPE), YOLOv11. |
| Subjects: | Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science) 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 > 36304-Automation Electronic Engineering |
| Depositing User: | Ismail Adrian Putra Yuniardi |
| Date Deposited: | 10 Aug 2026 01:58 |
| Last Modified: | 10 Aug 2026 01:58 |
| URI: | http://repository.its.ac.id/id/eprint/144179 |
Actions (login required)
 |
View Item |