Sistem Traffic Light Otomatis Berdasarkan Panjang Antrian Kendaraan dengan Pengolahan Citra
Automatic Traffic Light System Based on Vehicle Queue Length with Image Processing
DOI:
https://doi.org/10.35793/jti.v19i02.52144Abstract
Abstract — This research is motivated by problems with traffic light as traffic signs which in several cases and certain places have not been able to overcome the problem of traffic jams which continue to occur accompanied by the rate of growth and population density in Indonesia. The traffic light system that is generally used today is the Fixed Time Traffic Light Controller, which in reality is less effective for this problem. For this reason, this research developed an Automatic Traffic Light System Based on Vehicle Queue Length with Image Processing with the aim of helping reduce congestion that occurs at intersections, by processing information through a working system on a computer. By using Image Processing, Python, Yolo, Matlab and going through testing stages, it can be concluded that the system can run and detect vehicles on every road lane. From overall car detection on each road section in five tests , it was found that the systems accuracy rate for detecting cars was 81%, and motorbike detection was 54%.
Key words— Automatic Traffic Light; Fuzzy Mamdani; Image Processing; YOLO.
Abstrak — Penelitian ini dilatarblakangi oleh permasalahan pada traffic light sebagai rambu lalu lintas yang pada beberapa kasus dan tempat tertentu belum bisa mengatasi permasalahan macet yang terus terjadi diiringi oleh laju pertumbuhan dan kepadatan penduduk di Indonesia. Sistem lampu lalu lintas yang umumnya digunakan sekarang yaitu Fixed Time Traffic Light Controller, yang pada kenyataannya kurang efektif untuk permasalahan tersebut. untuk itu penelitian ini mengembangkan Sistem Traffic Light Otomatis Berdasarkan Panjang Antrian Kendaraan dengan Pengolahan Citra dengan tujuan untuk membantu mengurangi kemacetan yang terjadi dipersimpangan, dengan mengolah informasi melalui sistem kerja pada komputer. Dengan menggunakan Pengolahan Citra, Python, YOLO, Matlab dan melalui tahapan pengujian, dapat disimpulkan bahwa sistem dapat berjalan dan mendeteksi kendaraan pada setiap jalur jalan. Dari pendeteksian mobil secara keseluruhan pada setiap ruas jalan dalam lima kali pengujian didapatkan bahwa tingkat akurasi sistem mendeteksi mobil sebesar 81%, dan pendeteksian sepeda motor sebesar 54%.
Kata kunci —Fuzzy Mamdani; Lampu Lalu Lintas Otomatis; Pengolahan Citra; YOLO.
Downloads
Published
Issue
Section
License
Copyright (c) 2024 Pierre Tawalujan, Sherwin R. U. A. Sompie, Pinrolinvic D. K. Manembu
This work is licensed under a Creative Commons Attribution 4.0 International License.