Perhitungan Kecepatan Mobil Menggunakan Data Video dari Pesawat Tanpa Awak Berbasis Deep Learning

Siregar, Arie (2020) Perhitungan Kecepatan Mobil Menggunakan Data Video dari Pesawat Tanpa Awak Berbasis Deep Learning. Other thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 07211640000051-MuhammadArieLS-TA OK.pdf]
Preview
Text
07211640000051-MuhammadArieLS-TA OK.pdf

Download (48MB) | Preview

Abstract

Berdasarkan data dari BPS (Badan Pusat Statistik) jumlah mobil penumpang, mobil bis dan mobil barang di Indonesia sebanyak 25.525.876 unit pada tahun 2017. Jumlah mobil penumpang, mobil bis dan mobil barang tersebut terus bertambah dari tahun ke tahun dan menyebabkan kemacetan lalu lintas. Menurut Peraturan Menteri No. 14 tahun 2006 pasal 6 dimana indikator tingkat pelayanan adalah kecepatan lalu lintas (untuk jalan luar kota). Sejak tahun 2015 Direktorat Lalu Lintas Polda Metro Jaya menggunakan pesawat tanpa awak untuk memantau arus lalu lintas membuat lebih fleksibel dan mudah untuk memantau mobil di daerah-daerah yang tidak ada infrastruktur pendukungnya. Maka dari itu, perhitungan kecepatan mobil menggunakan pesawat tanpa awak dibutuhkan untuk membantu pemantauan arus lalu lintas di daerah-daerah yang tidak memiliki infrastruktur pendukung, ditambah biaya pemasangan dan pemeliharaan yang tidak besar. Hasil tangkapan kamera akan mendeteksi objek berbasis deep learning dengan metode YOLO (You Only Look Once), yang kemudian di tracking dengan metode SORT (Sort Online and Real-Time Tracking) sehingga membedakan tiap mobil lalu mobil yang melewati RoI (Region of Interest) akan dihitung kecepatannya. Dari hasil pengujian diperoleh pada RoI 5 meter program memiliki akurasi tertinggi di ketinggian 17 meter dengan nilai galat 4,74%, pada RoI 10 meter program memiliki akurasi tertinggi di ketinggian 17 meter dengan nilai galat 4,27% dan pada RoI 15 meter program memiliki akurasi tertinggi di ketinggian 17 meter dengan nilai galat 0,97%.
========================================================================================================================
Based on data from BPS (Central Statistics Agency) the number of passenger cars, bus cars and freight cars in Indonesia was 25,525,876 units in 2017. The number of passenger cars, bus cars and freight cars continues to increase from year to year and cause traffic congestion . According to Minister Regulation No. 14 of 2006 article 6 where the indicator of service level is traffic speed (for roads outside the city). Since 2015 the Traffic Directorate of the Jakarta Metropolitan Polda uses unmanned aircraft to monitor traffic flow making it more flexible and easier to monitor cars in areas where there is no supporting infrastructure. Therefore, the calculation of car speed using unmanned aircraft is needed to help monitor traffic flow in areas that do not have supporting infrastructure, plus installation and maintenance costs that are not large. The camera catch will detect objects based on deep learning with the YOLO (You Only Look Once) method, which is then tracked with the SORT (Sort Online and Real-Time Tracking) method so that it differentiates each car and then the car that passes through the RoI (Region of Interest) will the speed is calculated. From the test results obtained on the 5 meter RoI the program has the highest accuracy at an altitude of 17 meters with an error value of 4.74%, on a 10 meter RoI the program has the highest accuracy at an altitude of 17 meters with an error value of 4.27% and on a RoI 15 The meter program has the highest accuracy at an altitude of 17 meters with an error value of 0.97%.

Item Type: Thesis (Other)
Uncontrolled Keywords: Deep Learning, Pesawat Tanpa Awak, Kecepatan, YOLO ============================================================== Deep Learning, Drone, Speed, YOLO
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK5105.546 Computer algorithms
T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7888.3 Digital computers
Divisions: Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Computer Engineering > 90243-(S1) Undergraduate Thesis
Depositing User: Muhammad Arie Ladhika Siregar
Date Deposited: 24 Aug 2020 08:15
Last Modified: 26 May 2023 13:43
URI: http://repository.its.ac.id/id/eprint/79757

Actions (login required)

View Item View Item