Please use this identifier to cite or link to this item:
http://repository.ipb.ac.id/handle/123456789/166024| Title: | PENERAPAN KECERDASAN BUATAN PADA CCTV UNTUK IDENTIFIKASI PAKAIAN DINAS LAPANGAN (PDL) SITE ENGINEER DI PROYEK KONSTRUKSI |
| Other Titles: | Implementation of Artificial Intelligence on CCTV for Identifying Field Uniforms (PDL) of Site Engineers in Construction Projects |
| Authors: | Fathonah, Lathifunnisa KURNIA, AQIM TRY |
| Issue Date: | 2025 |
| Publisher: | IPB University |
| Abstract: | AQIM TRY KURNIA. Penerapan Kecerdasan Buatan pada CCTV untuk
Identifikasi Pakaian Dinas Lapangan (PDL) Site Engineer di Proyek Konstruksi.
Dibimbing oleh Lathifunnisa Fathonah, S.ST, M.T.
Penelitian ini mengembangkan sistem pemantauan keselamatan kerja
berbasis CCTV yang terintegrasi dengan Raspberry Pi 5 dan algoritma YOLOv8
untuk mendeteksi penggunaan Pakaian Dinas Lapangan (PDL) secara otomatis di
proyek konstruksi PT Wijaya Karya (Persero) Tbk. Model YOLOv8s dilatih
menggunakan 3.712 citra dari berbagai kondisi dan sudut pandang, menghasilkan
nilai mAP@0.5 sebesar 88,6%. Sistem mampu mengklasifikasikan individu ke
dalam empat kategori: site-engineer, buruh, unknown, dan person. Untuk
menjalankan model secara real-time di perangkat edge, Raspberry Pi dilengkapi
modul akselerator Hailo AI Kit (M.2 HAT) dan model dikonversi ke format .hef.
Saat terjadi pelanggaran penggunaan APD, sistem secara otomatis menangkap citra,
mengirim notifikasi melalui protokol MQTT, dan menampilkan data ke dashboard
pemantauan. Dengan sistem ini, proses pengawasan menjadi lebih cepat, konsisten,
dan tidak bergantung pada pemantauan manual, sekaligus mendukung penerapan
teknologi AI di lapangan secara langsung. AQIM TRY KURNIA. Penerapan Kecerdasan Buatan pada CCTV untuk Identifikasi Pakaian Dinas Lapangan (PDL) Site Engineer di Proyek Konstruksi. Supervised by Lathifunnisa Fathonah, S.ST, M.T. This study developed a workplace safety monitoring system using CCTV integrated with Raspberry Pi 5 and the YOLOv8 deep learning algorithm to automatically detect the use of Field Uniform (PDL) by site engineers at PT Wijaya Karya (Persero) Tbk construction projects. The YOLOv8s model was trained on 3.712 images under various conditions and perspectives, achieving a mAP@0.5 score of 88.6%. The system classifies individuals into four categories: site-engineer, buruh, unknown, and person. To enable real-time inference on edge devices, the Raspberry Pi is equipped with a Hailo AI Kit (M.2 HAT) and the model is converted to .hef format. When a PPE violation is detected, the system automatically captures the evidence image, sends a real-time notification via the MQTT protocol, and displays the data on a monitoring dashboard. This system enables faster, more consistent surveillance without manual intervention, supporting the direct application of AI in the field. |
| URI: | http://repository.ipb.ac.id/handle/123456789/166024 |
| Appears in Collections: | UT - Computer Engineering Tehcnology |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| cover_J0304211020_358792a5e1264341a1953f6afce2217c.pdf | Cover | 930.32 kB | Adobe PDF | View/Open |
| fulltext_J0304211020_5f66fe0c43ff4f69a19dca8d78214672.pdf Restricted Access | Fulltext | 2.19 MB | Adobe PDF | View/Open |
| lampiran_J0304211020_e4a9caaea2464dac95ca56e767f9449b.pdf Restricted Access | Lampiran | 264.75 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.