Please use this identifier to cite or link to this item: http://repository.ipb.ac.id/handle/123456789/165012
Title: Sistem Deteksi Hama Ulat Grayak (Spodoptera exigua) pada Bawang Merah (Allium cepa L.) Menggunakan Metode Deep Learning Berbasis Raspberry Pi
Other Titles: Detection System for Shallot (Allium cepa L.) Armyworm (Spodoptera exigua) Using Deep Learning Method Based Raspberry Pi
Authors: Pertiwi, Setyo
Utama, Deo Alif
Issue Date: 2025
Publisher: IPB University
Abstract: Peningkatan konsumsi bawang merah tidak selalu diiringi dengan peningkatan produksinya karena berbagai kendala yang dapat menurunkan produktivitas tanaman. Salah satu kendala yang sering dihadapi petani dalam produksi bawang merah adalah Organisme Pengganggu Tanaman (OPT). Penelitian ini bertujuan merancang dan menciptakan sistem deteksi hama ulat grayak untuk membantu proses monitoring hama dengan menerapkan salah satu model deep learning yaitu YOLOv8 untuk pendeteksian hama yang tampak pada daun tanaman bawang merah. Penelitian ini menggunakan metode deep learning dengan algoritma YOLOv8 yang terdiri dari tahapan: (1) pengumpulan dataset, (2) pengolahan dataset, (3) pelabelan dataset, (4) training dataset, (5) perancangan dan pembuatan sistem, (6) pengujian model dan sistem, serta (7) evaluasi. Berdasarkan hasil penelitian, nilai evaluasi menggunakan dataset testing, YOLOv8 menghasilkan model terbaik dalam melakukan deteksi hama dengan nilai akurasi 96,87%, nilai presisi 95,35%, nilai recall 92,71% dan nilai F1-Score 94,05%. Pengujian langsung pada tanaman bawang merah juga menunjukkan bahwa model dan sistem yang dikembangkan tidak hanya memiliki performa yang baik selama proses pelatihan, tetapi juga mampu mendeteksi keberadaan ulat grayak dengan baik pada saat implementasi nyata di lingkungan pertanian.
The increasing consumption of shallots is not always accompanied by a corresponding rise in production, primarily due to various obstacles that can reduce crop productivity. One of the major challenges faced by farmers in shallot cultivation is the presence of Plant Pests and Diseases (PPD), particularly the beet armyworm (Spodoptera exigua). This study aims to design and develop a pest detection system specifically for detecting beet armyworms, utilizing a deep learning approach based on the YOLOv8 algorithm to identify visible pests on shallot leaves. The research methodology includes the following stages: (1) dataset collection, (2) dataset preprocessing, (3) dataset labelling, (4) model training, (5) system design and development, (6) model and system testing, and (7) evaluation. Based on the results, the evaluation using the test dataset shows that the YOLOv8-based model achieved high performance in pest detection, with an accuracy of 96.87%, precision of 95.35%, recall of 92.71%, and an F1-score of 94.05%. Field tests conducted directly on shallot crops further demonstrated that the developed model and system not only performed well during training but were also effective in real-world implementation for detecting the presence of beet armyworms in agricultural environments.
URI: http://repository.ipb.ac.id/handle/123456789/165012
Appears in Collections:UT - Agricultural and Biosystem Engineering

Files in This Item:
File Description SizeFormat 
cover_F1401211032_0070183c81404e2486ed0ec6c1c8149f.pdfCover822.08 kBAdobe PDFView/Open
fulltext_F1401211032_3e22066dce09461aaceecb12da33a7b7.pdf
  Restricted Access
Fulltext3.87 MBAdobe PDFView/Open
lampiran_F1401211032_4f436b0cd3bc4f758495e82e6347ef30.pdf
  Restricted Access
Lampiran1 MBAdobe PDFView/Open


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.