IMPLEMENTASI METODE PEMILIHAN JENIS SAMPAH MENGGUNAKAN YOU ONLY LOOK ONCE (YOLO)

Marsha Safira, Adella (2024) IMPLEMENTASI METODE PEMILIHAN JENIS SAMPAH MENGGUNAKAN YOU ONLY LOOK ONCE (YOLO). Skripsi thesis, INSTITUT INFORMATIKA DAN BISNIS DARMAJAYA.

[img] Text
cover skripsi.pdf
Download (144kB)
[img] Text
Pernyataan Orisinalitas.pdf
Download (171kB)
[img] Text
halaman persetujuan.pdf
Download (212kB)
[img] Text
halaman pengesahan.pdf
Download (201kB)
[img] Text
Halaman Persembahan.pdf
Download (71kB)
[img] Text
abstrak.pdf
Download (514kB)
[img] Text
Halaman Moto.pdf
Download (36kB)
[img] Text
Riwayat Hidup.pdf
Download (62kB)
[img] Text
Kata Penghantar.pdf
Download (89kB)
[img] Text
Daftar Isi.pdf
Download (88kB)
[img] Text
Daftar Tabel.pdf
Download (74kB)
[img] Text
Daftar Gambar.pdf
Download (72kB)
[img] Text
Daftar Lampiran.pdf
Download (44kB)
[img] Text
Bab 1.pdf
Download (96kB)
[img] Text
Bab 2.pdf
Download (164kB)
[img] Text
Bab 3.pdf
Download (141kB)
[img] Text
Bab 4.pdf
Download (454kB)
[img] Text
Bab 5.pdf
Download (59kB)
[img] Text
Daftar Pustaka.pdf
Download (104kB)
[img] Text
Lampiran.pdf
Download (1MB)

Abstract

The waste problem is a major concern in various regions, including the Bandar Lampung City, which recorded daily waste accumulation of as much as 287,057.55 tons. Data from SIPSN KLHK in 2022 shows that the number of national waste pile reached 21.1 million tons. Previous research has classified recyclable waste used a Support Vector Machine and Local Binary Pattern. This research aims to develop a system implementation of separation of organic and non-organic waste using YOLOv8. Research methods include pre-training, data collection from sourced such as GitHub and Kaggle (3021 images in total), as well as manual labeling using Roboflow with the division of organic and non-organic labels. Preprocessing involves dividing the dataset into Train Set (70%), Valid Set (20%), and Test Set (10%). The YOLOv8 model is implemented and compared with other deep learning-based methods. Analyze device requirements including the hardware and software used, including specifications hardware and software such as YOLOv8, Google Colab, and Roboflow. Evaluation of research results including distribution of datasets, descriptions of research, and evaluation using the Roboflow API. The research results show that YOLOv8 is successful in detecting organic and non-organic waste, with effective dataset sharing. The use of GPUs and the Roboflow API makes a significant contribution to accelerating training and system integration. In conclusion, this research contributes to the development of technology for more efficient waste management in the Bandar Lampung City and the field of Computer Vision in general. Keywords: YOLOv8 Model, Deep Learning, Garbage Classification.

Item Type: Thesis (Skripsi)
Subjects: Ilmu Komputer
eSkripsi
Divisions: Skripsi/TA & PKPM/KP - Fakultas Ilmu Komputer > Teknik Informatika
Depositing User: Adella Marsha Safira
Date Deposited: 16 Apr 2024 05:48
Last Modified: 16 Apr 2024 05:48
URI: http://repo.darmajaya.ac.id/id/eprint/15359

Actions (login required)

View Item View Item