Muhammad Ghariza Pranaya Asari, . (2025) ANALISIS SISTEM DETEKSI REAL-TIME INDIKASI TINDAKAN KEKERASAN DAN SENJATA TAJAM/SENJATA API OTOMATIS MENGGUNAKAN MODEL YOLOV8. Skripsi thesis, Universitas Pembangunan Nasional Veteran Jakarta.
![[img]](http://repository.upnvj.ac.id/style/images/fileicons/text.png) |
Text
ABSTRAK.pdf Download (219kB) |
|
Text
AWAL.pdf Download (778kB) |
|
Text
BAB 1.pdf Restricted to Repository UPNVJ Only Download (342kB) |
|
Text
BAB 2.pdf Restricted to Repository UPNVJ Only Download (1MB) |
|
Text
BAB 3.pdf Restricted to Repository UPNVJ Only Download (562kB) |
|
Text
BAB 4.pdf Restricted to Repository UPNVJ Only Download (1MB) |
|
Text
BAB 5.pdf Download (262kB) |
|
Text
DAFTAR PUSTAKA.pdf Download (285kB) |
|
Text
RIWAYAT HIDUP.pdf Restricted to Repository UPNVJ Only Download (301kB) |
|
Text
LAMPIRAN.pdf Restricted to Repository UPNVJ Only Download (319kB) |
|
Text
HASIL PLAGIARISME.pdf Restricted to Repository staff only Download (20MB) |
|
Text
ARTIKEL KI.pdf Restricted to Repository staff only Download (535kB) |
Abstract
Violence in public spaces has become an increasingly alarming issue, especially in densely populated areas such as campus environments. Conventional surveillance systems using CCTV are limited in terms of detection efficiency and accuracy. This study aims to design and implement a real-time automated violence detection system using the YOLOv8 (You Only Look Once version 8) model. YOLOv8 was chosen for its ability to perform fast and accurate object detection, making it suitable for identifying violent actions as well as the presence of sharp weapons and firearms. The research process involved frame extraction from video, data annotation using Roboflow, model training with PyTorch, data augmentation, model fine-tuning, and system testing under various lighting and motion conditions. The evaluation results show that the optimized model achieved a mean Average Precision (mAP@0.5) of 0.902, an F1-score of 0.87, precision ranging from 91% to 95%, and recall of 90%, which significantly improved compared to the baseline model without optimization (mAP 0.813 and F1-score 0.76). Moreover, the system operates in real-time at an average speed of 21.95 frames per second (FPS) with a processing time of approximately 45.55 milliseconds per frame, tested on an NVIDIA RTX 2060 Laptop GPU. The system is also equipped with a graphical user interface (GUI) that provides real-time detection visualization and automatic event logging. This research demonstrates that the YOLOv8-based violence detection system delivers high performance and holds strong potential for real-world implementation in enhancing public security monitoring.
| Item Type: | Thesis (Skripsi) |
|---|---|
| Additional Information: | [No.Panggil: 2110511126] [Pembimbing 1: Didit Widiyanto] [Pembimbing 2: Nurul Afifah Arifuddin] [Penguji 1: Musthofa Galih Pradana] [Penguji 2: Novi Trisman Hadi] |
| Uncontrolled Keywords: | Computer vision, GUI, real-time, violence detections, weapons, YOLOv8 |
| Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76 Computer software |
| Divisions: | Fakultas Ilmu Komputer > Program Studi Informatika (S1) |
| Depositing User: | MUHAMMAD GHARIZA PRANAYA ASARI |
| Date Deposited: | 05 Aug 2025 08:22 |
| Last Modified: | 05 Aug 2025 08:22 |
| URI: | http://repository.upnvj.ac.id/id/eprint/37406 |
Actions (login required)
 |
View Item |