AKBAR, FIKRI TAUFIK (2023) RANCANG BANGUN SISTEM PENDETEKSI MASKER WAJAH BERBASIS NVIDIA JETSON NANO. S1 thesis, Fakultas Teknik Universitas Sultan Ageng Tirtayasa.
![[img]](https://eprints.untirta.ac.id/style/images/fileicons/text.png) |
Text (SKRIPSI)
Fikri Taufik Akbar_3332150043_Fulltext.pdf Restricted to Registered users only Download (3MB) |
|
Text (SKRIPSI)
Fikri Taufik Akbar_3332150043_01.pdf Restricted to Registered users only Download (1MB) |
|
Text (SKRIPSI)
Fikri Taufik Akbar_3332150043_02.pdf Restricted to Registered users only Download (145kB) |
|
Text (SKRIPSI)
Fikri Taufik Akbar_3332150043_03.pdf Restricted to Registered users only Download (450kB) |
|
Text (SKRIPSI)
Fikri Taufik Akbar_3332150043_04.pdf Restricted to Registered users only Download (1MB) |
|
Text (SKRIPSI)
Fikri Taufik Akbar_3332150043_05.pdf Restricted to Registered users only Download (60kB) |
|
Text (SKRIPSI)
Fikri Taufik Akbar_3332150043_Ref.pdf Restricted to Registered users only Download (132kB) |
|
Text (SKRIPSI)
Fikri Taufik Akbar_3332150043_Lamp.pdf Restricted to Registered users only Download (209kB) |
Abstract
This study designed a system that can detect masks, including wearing masks, not wearing masks and wearing masks but wrong. This system was created using a program in Python and NVIDIA Jetson Nano hardware. Mask detection using the TensorRT framework and the YoloV4 model. Previously, the YoloV4 model was trained to be able to perform mask detection. The training process is carried out at the Google Collaboratory because there is a super GPU. The results of the training process were then optimized so that the size of the YoloV4 model was reduced. After the YoloV4 model is created, the next step is to create a program to perform mask detection using NVIDIA Jetson Nano. If the system detects wearing a mask with a minimum detection amount of 99%, the buzzer will sound. The test was carried out using 10 masks with different types of masks and mask colors and with a total of 11 people, the average accuracy in the condition of wearing a mask was 94,742%, in the condition of not wearing a mask it was 92,026% and in the condition of wearing a mask but wrong. by 87,576%.
| Item Type: | Thesis (S1) | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Contributors: |
|
|||||||||
| Additional Information: | Penelitian ini merancang sebuah sistem yang dapat melakukan deteksi masker, diantaranya memakai masker, tidak memakai masker dan memakai masker tetapi salah. Sistem ini dibuat menggunakan program dengan bahasa Python dan perangkat keras NVIDIA Jetson Nano. Deteksi masker menggunakan framework TensorRT dan model YoloV4. Sebelumnya, model dari YoloV4 di-training agar dapat melakukan deteksi masker. Proses training dilakukan di Google Collaboratory karena terdapat super GPU. Hasil dari proses training, kemudian di optimalisasi agar ukuran model YoloV4 mengecil. Setelah model YoloV4 dibuat, maka selanjutnya membuat program untuk melakukan deteksi masker menggunakan NVIDIA Jetson Nano. Jika sistem mendeteksi memakai masker dengan jumlah deteksi minimal sebesar 99% maka buzzer akan berbunyi. Pengujian dilakukan dengan menggunakan 10 masker dengan tipe masker serta warna masker yang berbeda-beda dan dengan jumlah sebanyak 11 orang, rata-rata akurasi pada kondisi memakai masker sebesar 94,742%, pada kondisi tidak memakai masker sebesar 92,026% dan pada kondisi memakai masker tetapi salah sebesar 87,576%. | |||||||||
| Subjects: | T Technology > TA Engineering (General). Civil engineering (General) T Technology > TK Electrical engineering. Electronics Nuclear engineering |
|||||||||
| Divisions: | 03-Fakultas Teknik 03-Fakultas Teknik > 20201-Jurusan Teknik Elektro |
|||||||||
| Depositing User: | Mr Fikri Taufik Akbar | |||||||||
| Date Deposited: | 20 Sep 2023 09:29 | |||||||||
| Last Modified: | 20 Sep 2023 09:29 | |||||||||
| URI: | http://eprints.untirta.ac.id/id/eprint/29768 |
Actions (login required)
 |
View Item |