KAMERA CERDAS BERBASIS COMPUTER VISION UNTUK DETEKSI PENGGUNAAN SAFETY HELMET DI AREA KERJA

CHRISTOPHORUS MICHAEL ROBIN CARAKA YUDA, . (2025) KAMERA CERDAS BERBASIS COMPUTER VISION UNTUK DETEKSI PENGGUNAAN SAFETY HELMET DI AREA KERJA. Sarjana thesis, UNIVERSITAS NEGERI JAKARTA.

	Text Cover.pdf Download (766kB)
	Text BAB I.pdf Download (222kB)
	Text BAB II.pdf Restricted to Registered users only Download (579kB) | Request a copy
	Text BAB III.pdf Restricted to Registered users only Download (4MB) | Request a copy
	Text BAB IV.pdf Restricted to Registered users only Download (755kB) | Request a copy
	Text BAB V.pdf Restricted to Registered users only Download (203kB) | Request a copy
	Text DAFTAR PUSTAKA.pdf Download (241kB)
	Text LAMPIRAN.pdf Restricted to Registered users only Download (3MB) | Request a copy

Abstract

Penelitian ini mengembangkan aplikasi deteksi pelanggaran helm pengaman berbasis YOLOv8 (format ONNX) untuk pemantauan keselamatan kerja secara real-time. Aplikasi dibuat dengan Python, menggunakan antarmuka Tkinter, dan terhubung ke dasbor web Flask untuk menampilkan log pelanggaran. Sistem dapat menerima input video dari webcam, kamera USB, atau kamera IP, dan mengklasifikasikan jenis helm yang digunakan. Hanya helm oranye dan putih yang dianggap valid, sementara kondisi lain memicu alarm dan menyimpan gambar bukti. Model YOLOv8 dilatih pada enam kelas dengan hasil presisi 0,921, recall 0,859, mAP50 0,919, dan mAP50-95 0,619. Evaluasi menunjukkan sistem stabil dan akurat dalam pemantauan otomatis berbasis visi komputer. ****** The application of computer vision technology in automation systems plays a vital role in improving the efficiency of occupational safety monitoring within industrial environments. This study develops a visual detection application based on the YOLOv8 model, converted into the ONNX format, to identify safety helmet violations in real time. The system is developed using Python with a Tkinter-based user interface and integrated with a Flask web dashboard that displays violation log data. The application accepts video input from various sources, including webcams, USB cameras, and IP cameras, to classify the types of helmets being worn. Only safety helmets in orange and white colors are considered valid. The detection of bare heads, motorcycle helmets, or helmets in noncompliant colors triggers an alarm and saves an image as evidence of the violation. The YOLOv8 model was trained on a six-class dataset and demonstrated strong performance, with a precision of 0.921, recall of 0.859, mAP50 of 0.919, and mAP50-95 of 0.619. System evaluation confirms the application's stability and accuracy for automated monitoring based on computer vision.

Item Type:	Thesis (Sarjana)
Additional Information:	1). Syufrijal, S.T.,M.T. ; 2). Ir. Heri Firmansyah, S.T.
Subjects:	Teknologi dan Ilmu Terapan > Teknik Elektronika
Divisions:	FT > D IV Teknologi Rekayasa Otomasi
Depositing User:	Christophorus Michael Robin Caraka Yuda .
Date Deposited:	15 Aug 2025 07:06
Last Modified:	15 Aug 2025 07:06
URI:	http://repository.unj.ac.id/id/eprint/60621

Actions (login required)

View Item