Claim Missing Document
Check
Articles

Found 2 Documents
Search

Deteksi Helm Pengendara dan Plat Nomor Kendaraan pada CCTV Lampu Lalu Lnitas Menggunakan Algoritma YOLO Muhammad Priyo Anugrah; Bagus Fatkhurrozi; Hery Teguh Setiawan
Voteteknika (Vocational Teknik Elektronika dan Informatika) Vol 12, No 1 (2024): Vol. 12, No 1, Maret 2024
Publisher : Universitas Negeri Padang

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24036/voteteknika.v12i1.122991

Abstract

Perlengkapan berlalu lintas merupakan hal yang wajib digunakan pada saat ini. Keselamatan menjadi alasan utama diharuskannya pengendara mengenakan perlengkapan berkendara khususnya helm. Namun masih banyak pengendara sepeda motor yang melanggar peraturan berlalulintas dengan tidak menggunakan helm saat berkendara. Perlu adanya sistem yang dapat mengawasi pengendara yang tidak mengenakan helm saat mengendarai sepeda motor. Hal ini bisa diterapkan menggunakan salah satu algoritma deteksi objek yaitu YOLO yang menggunakan Bahasa pemograman Python. YOLO digunakan sebagai alat yang akan menendeteksi dan mengenali objek seperti pengendara sepeda motor yang tidak menggunakan helm saat berkendara dari CCTV lampu lalulitas. Penelitian ini  dilakukan dengan harapan peneliti mengentahui apakah algoritma deteksi objek seperti yolov5 bisa dimanfaaatkan sebagai alat untuk mendeteksi pelanggaran berkandara seperti helm dan menangkap gambar pelanggar apabila kamera diambil dari CCTV lampu lalulintas. Penelitian ini akan akan memiliki 2 tahap penelitian yaitu, tahap pelatihan dataset dan tahap pengujian hasil training. Pada tahap pelatihan berfokus dalam melakukan training custom data untuk algoritma yolo dan tahap pengujian dilakukan untuk menguji apakah algoritma yolo yang telah dilakukan pelatihan sesuai dengan yang di harapkan. Hasil pengujian didapatkan nilai akurasi rata-rata dari semua class objek yang dipakai sebesar 94.52% dan error yang dihasilka oleh aplikasi yang dibuat adalah sebesar 4,66%.Kata kunci : CCTV, Deteksi objek, helm, Python, YOLOv5. The use of traffic safety equipment is now mandatory. Safety is the main reason why riders wear safety equipment, especially helmets. However, there are still many motorcyclists who violate traffic regulations by not wearing helmets while riding. Therefore, a system is needed that can monitor riders who do not wear helmets while riding motorcycles. This system can be implemented using one of the object detection algorithms, YOLO, which uses the Python programming language. YOLO can be used as a tool to detect and recognize objects, such as motorcyclists who do not use helmets while driving, through CCTV traffic lights. This research was conducted with the aim to find out whether object detection algorithms, such as YOLOv5, can be utilized as a tool to detect driving violations, such as not wearing a helmet, and capture images of violators through traffic light CCTV cameras. This research consists of 2 stages, namely the dataset training stage and the training result testing stage. The training stage focuses on training data specific to the YOLO algorithm, while the testing stage aims to test whether the YOLO algorithm that has been trained is as expected. The results of testing the application show that the average accuracy value of all object classes used is 94.52%, and the error generated by the application is 4.66%.Keywords: CCTV, Object detection, helmet, YOLOv5, Python.
Performance Evaluation of Overcurrent Relay Coordination on 20-kV Busbar and Feeders Ageng Rizkianto; Agung Trihasto; Deria Pravitasari; Hery Teguh Setiawan
Aviation Electronics, Information Technology, Telecommunications, Electricals, and Controls (AVITEC) Vol 8, No 1 (2026): February
Publisher : Institut Teknologi Dirgantara Adisutjipto

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.28989/avitec.v8i1.3484

Abstract

The coordination of Overcurrent Relays (OCR) in power systems is crucial to ensure selectivity and reliability. Mis-coordination between OCRs on the 20-kV busbar and feeders can significantly reduce system performance, often due to improper determination of pick-up values and Time Multiplier Setting (TMS). Previous studies mostly focused on protection coordination for a single feeder and relied solely on simulation. This study evaluates OCR coordination on the 20-kV busbar and five feeders connected to the Unit I transformer at Secang Substation by combining manual analysis and Electrical Transient Analyzer Program (ETAP) simulations, validated against IEEE Std 242-2001. This integrated approach provides more reliable insights than earlier works limited to single-feeder coordination or software-only methods. Evaluation was conducted through short-circuit current analysis and Time Current Characteristic (TCC) curves, yielding pick-up and TMS values that produce Coordination Time Intervals (CTI) in compliance with IEEE Std 242-2001. Results indicate that the busbar OCR achieved a pick-up of 0.566 and a TMS of 0.236. For the feeders, SCG 10 achieved 0.27 and 0.173; SCG 03 yielded 0.5025 and 0.147; SCG 05 produced 0.441 and 0.153; SCG 07 yielded 0.35 and 0.165; and SCG 08 achieved 0.5535 and 0.137. Applying these settings produced CTI values exceeding the minimum requirement of 0.3 seconds. This evaluation demonstrates that coordinated OCR settings can improve reliability in 20-kV distribution systems and reduce the risk of widespread outages due to protection failures.