Heryansyah, Nova (2025) Sistem Deteksi Objek untuk Identifikasi Jenis Beras Menggunakan Algoritma YoloV8. Undergraduate thesis, Universitas Pembangunan Jaya.
![[img]](https://eprints.upj.ac.id/style/images/fileicons/text.png) |
Text
Pendahuluan(REV).pdf Download (1MB) |
|
Text
Abstract.pdf Download (17kB) |
|
Text
Abstrak.pdf Download (16kB) |
|
Text
Daftar Isi.pdf Download (471kB) |
|
Text
Daftar Gambar.pdf Download (14kB) |
|
Text
Daftar Tabel.pdf Download (15kB) |
|
Text
Daftar Lampiran.pdf Download (305kB) |
|
Text
BAB I.pdf Download (43kB) |
|
Text
BAB II.pdf Download (63kB) |
|
Text
BAB III.pdf Download (63kB) |
|
Text
BAB IV.pdf Download (186kB) |
|
Text
BAB V.pdf Download (292kB) |
|
Text
BAB VI.pdf Download (26kB) |
|
Text
Daftar Pustaka.pdf Download (280kB) |
|
Text
Lampiran.pdf Download (2MB) |
|
Text
Berita Acara Unggah Mandiri.pdf Download (297kB) |
|
Text
Bukti Lolos Similarity (Plagiarism Org).pdf Download (498kB) |
Abstract
Manual identification of rice varieties is often inconsistent due to limited visual knowledge and high inter-varietal similarity. This research aims to develop an object detection system based on the YOLOv8 algorithm that can automatically and in real-time identify different rice varieties. The system was trained using an image dataset consisting of seven rice types: black rice, IR42, sticky rice, red rice, basmati, buloq, and japonica. The training process produced a classification model with high accuracy on internal data. Evaluation was conducted using confusion matrix, black box, and white box testing. The confusion matrix results showed 99% accuracy on internal data, while accuracy on external (outsourced) data dropped to 82%. Black box testing confirmed that the system functioned as expected within the user interface. White box testing was performed by analyzing the program flow and decision-making process in the source code. The decrease in performance on external data was attributed to the lack of diversity in the training data, suggesting the need for data augmentation, additional images, or re-training. The developed system demonstrates promising potential as an efficient solution for rice variety classification to support digitalization in the agricultural sector.
| Item Type: | Karya Tulis Ilmiah (KTI) (Undergraduate) |
|---|---|
| Uncontrolled Keywords: | deteksi objek; YOLOv8; computer vision; deep learning; teknologi pertanian |
| Subjects: | L Education > L Education (General) T Technology > T Technology (General) Z Bibliography. Library Science. Information Resources > Z719 Libraries (General) Z Bibliography. Library Science. Information Resources > ZA Information resources > ZA4050 Electronic information resources Z Bibliography. Library Science. Information Resources > ZA Information resources > ZA4450 Databases |
| Divisions: | Fakultas Teknologi dan Desain > Informatika |
| Depositing User: | Nova Heryansyah |
| Date Deposited: | 08 Aug 2025 03:18 |
| Last Modified: | 08 Aug 2025 03:21 |
| URI: | http://eprints.upj.ac.id/id/eprint/11747 |
Actions (login required)
 |
View Item |