Deteksi Lampu Rem Dan Lampu Sein Pada Mobil Otonom Menggunakan Metode YOLOv5

Wijayanto, Kivlan Khair (2023) Deteksi Lampu Rem Dan Lampu Sein Pada Mobil Otonom Menggunakan Metode YOLOv5. Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 07111940000170-Undergraduate_Thesis.pdf] Text
07111940000170-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 October 2025.

Download (1MB) | Request a copy

Abstract

Lampu Sein dan Lampu rem merupakan komponen penting pada kendaraan bermotor karena memiliki peran penting dalam menjaga keamanan saat berkendara. Dengan perkembangan yang pesat pada mobil otonom, diperlukan untuk mobil otonom mendeteksi lampu sein dan lampu rem. Pada penelitian ini, metode object detection menggunakan metode CNN (Convolutional Neural Network) pada model algoritma YOLOv5. Penulis menggunakan model algoritma YOLOv5 untuk membedakan penggunaan lampu sein dan lampu rem berdasarkan akurasi, dan kecepatan deteksi objek. Penelitian dilakukan dengan pengambilan gambar mobil pada jalan, diaplikasikan Wiener Filter untuk mengurangi noise, pembagian gambar untuk data train, validation dan test, evaluasi metrik, dan deteksi objek. Berdasarkan evaluasi metrik, model YOLOv5l memiliki nilai akurasi tertinggi dengan nilai 0.811. Namun untuk kecepatan deteksi tertinggi dimiliki model YOLOv5n dan YOLOv5s dengan nilai 106 FPS. Model dengan nilai akurasi tinggi dengan kecepatan deteksi objek tinggi dimiliki model YOLOv5l dengan akurasi 0.848 dan 47.3 FPS. Model algoritma YOLOv5 dapat melakukan deteksi lampu sein dan lampu rem.
=================================================================================================================================
Turn signal lights and brake lights are important components in motor vehicles because they play a crucial role in ensuring safety while driving. With the rapid development of autonomous cars, it is necessary for autonomous vehicles to detect turn signal lights and brake lights. In this study, the object detection method using the CNN (Convolutional Neural Network) algorithm on YOLOv5 was employed. The author used the YOLOv5 algorithm to differentiate between the usage of turn signal lights and brake lights based on accuracy and object detection speed. The research was conducted by capturing images of cars on the road, applying Wiener Filter to reduce noise, dividing the images into training, validation, and test data, evaluating metrics, and object detection. Based on the evaluation metrics, the YOLOv5x model had the highest accuracy value of 0.811. However, the models YOLOv5n and YOLOv5s had the highest detection speed of 106 FPS. The model with high accuracy and high object detection speed is YOLOv5l with an accuracy of 0.848 and 47.3FPS.

Item Type: Thesis (Other)
Uncontrolled Keywords: Lampu Sein, Lampu Rem, Mobil Otonom, YOLOv5, Autonomous Cars, Turn Signal Lights and Brake Lights Detection,
Subjects: Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines.
Q Science > QA Mathematics > QA336 Artificial Intelligence
T Technology > TL Motor vehicles. Aeronautics. Astronautics > TL152.8 Vehicles, Remotely piloted. Autonomous vehicles.
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User: Kivlan Khair Wijayanto
Date Deposited: 20 Nov 2023 01:18
Last Modified: 20 Nov 2023 01:18
URI: http://repository.its.ac.id/id/eprint/102243

Actions (login required)

View Item View Item