Pengenalan Rambu Lalu Lintas Menggunakan Metode Feature Fusion Single Shot Multibox Detector (FSSD)

Saputri, Dinda Ayu (2022) Pengenalan Rambu Lalu Lintas Menggunakan Metode Feature Fusion Single Shot Multibox Detector (FSSD). Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 06111740000015-Undergraduate_Thesis.pdf] Text
06111740000015-Undergraduate_Thesis.pdf - Accepted Version
Restricted to Repository staff only until 1 April 2024.

Download (5MB) | Request a copy

Abstract

Kemajuan teknologi saat ini mendorong manusia untuk mengembangkan berbagai sarana terutama di bidang transportasi. Intelligent Transportation System (ITS) merupakan penerapan teknologi canggih yang mampu memanajemen lalu lintas. Advanced Driver Assistance Systems (ADAS) adalah penerapan ITS pada kendaraan. ADAS memiliki banyak sistem yang dirancang untuk memberikan keamanan dan kenyamanan yang lebih baik dalam berkendara. Salah satu sistem dari ADAS, Traffic Sign Recognition (TSR) merupakan sebuah sistem untuk melakukan pengenalan terhadap rambu lalu lintas. Penelitian Tugas Akhir ini bertujuan untuk merancang sebuah sistem pengenalan rambu-rambu lalu lintas pada video rekaman perjalanan saat berkendara, rambu dideteksi kemudian diidentifikasi sesuai label jenis rambu dengan menggunakan metode Feature Fusion Single Shot Multibox Detector (FSSD) Alur pendeteksian rambu lalu lintas ini diantaranya input video, input RoI, proses FSSD. Uji coba penelitian dilakukan dengan dengan dua variasi skenario pengujian yaitu uji coba berdasarkan kuantitas dan posisi rambu serta uji coba berdasarkan kecepatan saat berkendara . Hasil penelitian menunjukan rata-rata akurasi pada salah satu uji coba mampu mencapai 93.66%. Nilai rata-rata akurasi sistem secara keseluruhan sebesar 87.52%.
================================================================================================
Current technological advances encourage humans to develop various means, especially in the field of transportation. Intelligent Transportation System (ITS) is the application of advanced technology capable of managing traffic. Advanced Driver Assistance Systems (ADAS) is the application of ITS to vehicles. ADAS has many systems designed to provide greater safety and comfort while driving. One of the ADAS systems, Traffic Sign Recognition (TSR) is a system for recognizing traffic signs. This final project research aims to design a traffic sign recognition system on video recordings of trips while driving, the signs are detected and then identified according to the type of sign label using the Feature Fusion Single Shot Multibox Detector (FSSD) method.This traffic sign detection flow includes video input, RoI input, FSSD process. The research trial was carried out with two variations of test scenarios, namely a test based on the quantity and position of the signs and a test based on speed when driving. The results showed that the average accuracy in one of the trials was able to reach 93.66%. The average value of the overall system accuracy is 87.52%.

Item Type: Thesis (Other)
Uncontrolled Keywords: Traffic Sign Recognition (TSR), Feature Fusion Single Shot Multibox Detector (FSSD), Deep Learning, RoI
Subjects: Q Science
Q Science > Q Science (General)
Q Science > QA Mathematics
Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions: Faculty of Science and Data Analytics (SCIENTICS) > Mathematics > 44201-(S1) Undergraduate Thesis
Depositing User: Dinda Ayu Saputri
Date Deposited: 08 Feb 2022 03:44
Last Modified: 02 Nov 2022 02:30
URI: http://repository.its.ac.id/id/eprint/93089

Actions (login required)

View Item View Item