SISTEM PEMANTAU DEMONSTRASI BERBASIS YOU ONLY LOOK ONCE VERSI 5 NANO (YOLOV5N) MENGGUNAKAN RASPBERRY PI

Maulana Ridhwan Riziq, . (2025) SISTEM PEMANTAU DEMONSTRASI BERBASIS YOU ONLY LOOK ONCE VERSI 5 NANO (YOLOV5N) MENGGUNAKAN RASPBERRY PI. Skripsi thesis, Universitas Pembangunan Nasional Veteran Jakarta.

[img] Text
ABSTRAK.pdf

Download (288kB)
[img] Text
AWAL.pdf

Download (689kB)
[img] Text
BAB 1.pdf
Restricted to Repository UPNVJ Only

Download (340kB)
[img] Text
BAB 2.pdf
Restricted to Repository UPNVJ Only

Download (864kB)
[img] Text
BAB 3.pdf
Restricted to Repository UPNVJ Only

Download (485kB)
[img] Text
BAB 4.pdf
Restricted to Repository UPNVJ Only

Download (1MB)
[img] Text
BAB 5.pdf

Download (307kB)
[img] Text
DAFTAR PUSTAKA.pdf

Download (303kB)
[img] Text
RIWAYAT HIDUP.pdf
Restricted to Repository UPNVJ Only

Download (204kB)
[img] Text
LAMPIRAN.pdf
Restricted to Repository UPNVJ Only

Download (763kB)
[img] Text
HASIL PLAGIARISME.pdf
Restricted to Repository staff only

Download (10MB)
[img] Text
ARTIKEL KI.pdf
Restricted to Repository staff only

Download (446kB)

Abstract

Early detection of crowd related hazards during protests is crucial for enabling rapid decision making by security personnel. This study develops a real time object detection system based on You Only Look Once (YOLO) to count crowds and detect fire from a top down view. Three lightweight variants—YOLOv3 Tiny, YOLOv4 Tiny, and YOLOv5n—were compared using precision, recall, F1 score, and mean average precision (mAP), as well as inference performance on a Raspberry Pi 4B. Evaluation results show that YOLOv5n outperforms the others with a precision of 91.5 %, recall of 91.6 %, F1 score of 91.5 %, and mAP@50 of 95.6 %, and was therefore selected for deployment. The system integrates a web interface that streams detection output in real time and issues alerts when crowd size exceeds a predefined threshold or when fire is detected. In tests using outdoor protest videos, the Raspberry Pi 4B running YOLOv5n achieved an average latency of 216 ms, jitter of 34.8 ms, throughput of 5.2 FPS, CPU utilization of 71.3 %, memory usage of 15.9 %, and an average confidence score of 65.3 %.

Item Type: Thesis (Skripsi)
Additional Information: [No. Panggil: 2110314033] [Pembimbing 1: Muhamad Alif Razi] [Pembimbing 2: Ni Putu Devira Ayu Martini] [Penguji 1: Achmad Zuchriadi P] [Penguji 2: Fajar Rahayu Ikhwannul Mariati]
Uncontrolled Keywords: YOLOv5n, Raspberry Pi 4B, Crowd Monitoring, Fire Detection, Real-Time Monitoring
Subjects: Q Science > Q Science (General)
Q Science > QA Mathematics > QA75 Electronic computers. Computer science
Q Science > QA Mathematics > QA76 Computer software
Divisions: Fakultas Teknik > Program Studi Teknik Elektro (S1)
Depositing User: MAULANA RIDHWAN RIZIQ
Date Deposited: 28 Jul 2025 06:38
Last Modified: 28 Jul 2025 06:38
URI: http://repository.upnvj.ac.id/id/eprint/38387

Actions (login required)

View Item View Item