RAFLI, RAFLI (2024) Deteksi dan Perhitungan Kendaraan Menggunakan YOLOv5 dan DeepSORT Pada Jalan Protokol Kota Cilegon. S1 thesis, Fakultas Teknik Universitas Sultan Ageng Tirtayasa.
![[img]](https://eprints.untirta.ac.id/style/images/fileicons/text.png) |
Text (SKRIPSI)
Rafli_3332190067_Fulltext.pdf Restricted to Registered users only Download (2MB) |
|
Text
Rafli_3332190067_01.pdf Restricted to Registered users only Download (2MB) |
|
Text
Rafli_3332190067_02.pdf Restricted to Registered users only Download (2MB) |
|
Text
Rafli_3332190067_03.pdf Restricted to Registered users only Download (2MB) |
|
Text
Rafli_3332190067_04.pdf Restricted to Registered users only Download (2MB) |
|
Text
Rafli_3332190067_05.pdf Restricted to Registered users only Download (2MB) |
|
Text
Rafli_3332190067_Ref.pdf Restricted to Registered users only Download (2MB) |
|
Text
Rafli_3332190067_Lamp.pdf Restricted to Registered users only Download (2MB) |
|
Text
Rafli_3332190067_CP.pdf Restricted to Registered users only Download (13MB) |
Abstract
Increase in human mobility by vehicle leading to traffic congestion which impacting human wellness. One of the crucial methods used in traffic management, vehicle counting is one of the keys in decreasing traffic congestion. Object detection techniques employing computer vision in vehicle counting methods outperform conventional counting systems. The simultaneous process of vehicle classification and counting with simplicity in deployment becomes a strong point of computer vision employment in this field. YOLOv5 algorithm usage in this research overcomes computational load issues in edge computing deployment combined with the DeepSORT tracker algorithm to increase the robustness of vehicle detection and counting. The overall average vehicle counting accuracy result in 12 real-time vehicle counting experiment obtained average accuracy of 72,61% for “Mobil” object class, 21,56% for “Motor” object class, 70% for “Bus” object class, 35,63% for “Truk” object class. The overall vehicle counting accuracy is 49,95%.
| Item Type: | Thesis (S1) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
|||||||||
| Additional Information: | Peningkatan mobilitas manusia menggunakan kendaraan menyebabkan kepadatan lalu lintas yang berdampak buruk pada kesehatan manusia. Sebagai salah satu metode utama yang digunakan dalam sistem manajemen lalu lintas, perhitungan kendaraan menjadi kunci dalam mengurai kemacetan. Pemanfaatan teknik deteksi objek menggunakan visi komputer dalam metode perhitungan kendaraan dapat mengatasi kelemahan sistem perhitungan secara konvensional. Proses klasifikasi jenis dan perhitungan kendaraan yang dilakukan secara bersamaan dengan implementasi yang sederhana menjadi keunggulan pemanfaatan visi komputer dalam perhitungan kendaraan. Penggunaan Algoritma YOLOv5 pada penelitian ini mengatasi permasalahan beban komputasi pada komputasi tepi yang dikombinasikan dengan algoritma DeepSORT untuk meningkatkan kekokohan deteksi dan perhitungan kendaraan. Hasil pengujian perhitungan kendaraan pada 12 pengujian secara langsung memiliki nilai akurasi rata-rata 72,61% untuk kelas “Mobil,” 21,56% untuk kelas “Motor,” 70% untuk kelas “Bus,” dan 35,63% untuk kelas “Truk.” Secara keseluruhan, nilai akurasi perhitungan adalah 49,95%. | |||||||||
| Subjects: | Q Science > QA Mathematics > QA75 Electronic computers. Computer science Q Science > QA Mathematics > QA76 Computer software |
|||||||||
| Divisions: | 03-Fakultas Teknik 03-Fakultas Teknik > 20201-Jurusan Teknik Elektro |
|||||||||
| Depositing User: | Mr Rafli Rafli | |||||||||
| Date Deposited: | 22 Mar 2024 08:51 | |||||||||
| Last Modified: | 22 Mar 2024 08:51 | |||||||||
| URI: | http://eprints.untirta.ac.id/id/eprint/34435 |
Actions (login required)
 |
View Item |