Please use this identifier to cite or link to this item:
http://repository.ipb.ac.id/handle/123456789/152750
Title: | Deteksi Penyakit Bercak Daun Alternaria dan Busuk Hitam pada Kembang Kol (Brassica oleracea var. botrytis) Berbasis Deep Learning |
Other Titles: | Detection of Alternaria Leaf Spot and Black Rot Diseases in Cauliflower (Brassica oleracea var. botrytis) Based on Deep Learning |
Authors: | Liyantono, Liyantono Giyanto, Giyanto Rijaldi, Rizki Moch |
Issue Date: | 2024 |
Publisher: | IPB University |
Abstract: | Produksi kembang kol pada tahun 2022 menurun sebesar 5,54% dari produksi
tahun sebelumnya. Penurunan produksi dapat disebabkan oleh berbagai faktor,
salah satunya adalah serangan hama dan penyakit. Deteksi terhadap penyakit
tertentu serta tindakan pengendalian yang tepat oleh petani sangat diperlukan untuk
mencegah kerugian yang ditimbulkan. Tujuan dari penelitian ini adalah membuat
model deep learning untuk deteksi penyakit bercak daun alternaria dan busuk hitam
pada kembang kol dan mengintegrasikannya dengan website sebagai sistem pakar.
Tahapan penelitian terdiri dari (1) pengumpulan dataset, (2) augmentasi dataset,
(3) pelabelan dataset, (4) training model, (5) evaluasi model, dan (6) implementasi
model. Dataset terdiri dari 3 kelas yaitu bercak daun alternaria, busuk hitam, dan
sehat. Hasil evaluasi model pada 354 gambar uji menunjukkan bahwa YOLOv8
memiliki accuracy 99%, precision 98,89%, recall 98,60%, dan F1-score 98,74%.
SSD MobileNetV2 memiliki accuracy 96,24%, precision 94,11%, recall 91,73%,
dan F1-score 92,91%. Dengan demikian, YOLOv8 lebih baik dalam mendeteksi
penyakit kembang kol dan telah mampu mendeteksi penyakit kembang kol dengan
akurasi yang tinggi. Model dapat digunakan untuk mendeteksi penyakit kembang
kol melalui website yang dikembangkan menggunakan HTML, CSS, dan Flask
yang dapat diakses secara publik. The production of cauliflower in 2022 decreased by 5.54% compared to the previous year. This decline in production can be attributed to various factors, one of which is pest and disease infestation. Detection of specific diseases and appropriate control measures by farmers are crucial to prevent resulting losses. The objective of this research is to develop a deep learning model for detecting alternaria leaf spot and black rot diseases in cauliflower and integrate it into a website as an expert system. The research phases consist of (1) dataset collection, (2) dataset augmentation, (3) dataset labeling, (4) model training, (5) model evaluation, and (6) model implementation. The dataset comprises three classes: alternaria leaf spot, black rot, and healthy. The evaluation results of the model on 354 test images show that YOLOv8 achieves an accuracy of 99%, precision of 98,89%, recall of 98,60%, and F1-score of 98,74%. SSD MobileNetV2 achieves an accuracy of 96,24%, precision of 94,11%, recall of 91,73%, and F1-score of 92,91%. Therefore, YOLOv8 is superior in detecting cauliflower diseases and has demonstrated high accuracy in detecting cauliflower diseases. The model can be used to detect cauliflower diseases through a website developed using HTML, CSS, and Flask, which can be executed on a local The model can be used to detect cauliflower diseases through a website developed using HTML, CSS, and Flask, which is publicly accessible. |
URI: | http://repository.ipb.ac.id/handle/123456789/152750 |
Appears in Collections: | UT - Agricultural and Biosystem Engineering |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Cover.pdf Restricted Access | Cover | 503.8 kB | Adobe PDF | View/Open |
Full Teks.pdf Restricted Access | Fulltext | 3.02 MB | Adobe PDF | View/Open |
Lampiran.pdf Restricted Access | Lampiran | 3.2 MB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.