Please use this identifier to cite or link to this item: https://repositori.uma.ac.id/handle/123456789/30262
Title: Pendekatan Arsitektur ResNet-34 Pada Convolutional Neural Network (CNN) Untuk Mengklasifikasi Penyakit Pada Tanaman Daun Kopi
Other Titles: ResNet-34 Architectural Approach in Convolutional Neural Networks (CNN) for Classifying Diseases in Coffee Leaves
Authors: Situmorang, Cessy Cindi Regina
metadata.dc.contributor.advisor: Sulilawati
Keywords: Convolutional Neural Network (CNN);ResNet-34;Klasifikasi Citra;Penyakit Daun Kopi;Deep Learning;Image Classification;Coffee Leaf Diseases
Issue Date: Apr-2026
Publisher: Universitas Medan Area
Series/Report no.: NPM;228160027
Abstract: Penyakit pada daun kopi merupakan salah satu faktor utama yang dapat menurunkan produktivitas tanaman kopi secara signifikan. Identifikasi penyakit secara manual membutuhkan keahlian khusus dan berpotensi menimbulkan kesalahan, sehingga diperlukan sistem yang mampu melakukan klasifikasi secara cepat dan akurat. Penelitian ini bertujuan untuk mengembangkan model klasifikasi penyakit daun kopi menggunakan metode Convolutional Neural Network (CNN) dengan arsitektur ResNet-34. Dataset yang digunakan adalah Ethiopian Coffee Leaf Disease Dataset yang diperoleh dari Kaggle, terdiri dari 11.130 citra daun kopi dengan empat kelas, yaitu Healthy, Leaf Rust, Phoma, dan Cercospora. Tahapan penelitian meliputi preprocessing data, augmentasi citra, serta pembagian dataset menjadi data training, validation, dan testing. Model dilatih menggunakan hyperparameter dengan epoch sebesar 50, batch size 32, optimizer Adam, dan learning rate 0.0001. Hasil evaluasi menunjukkan bahwa model ResNet-34 menghasilkan akurasi training sebesar 99,44%, validation accuracy sebesar 98,56%, dan akurasi testing sebesar 98,83%. Seluruh kelas penyakit berhasil diklasifikasikan dengan nilai F1-Score di atas 0.97, dengan kelas Phoma mencapai nilai tertinggi sebesar 0.9991. Model selanjutnya diimplementasikan ke dalam aplikasi mobile untuk mendukung deteksi penyakit daun kopi secara real-time. Hasil penelitian ini membuktikan bahwa arsitektur ResNet-34 efektif dalam mengklasifikasikan penyakit daun kopi dan berpotensi dikembangkan lebih lanjut sebagai sistem pendukung pertanian digital. Coffee leaf disease is a major factor that can significantly reduce coffee plant productivity. Manual disease identification requires specialized expertise and is prone to errors, necessitating an automated system capable of fast and accurate classification. This study aims to develop a coffee leaf disease classification model using a Convolutional Neural Network (CNN) with the ResNet-34 architecture. The dataset used was the Ethiopian Coffee Leaf Disease Dataset obtained from Kaggle, consisting of 11,130 coffee leaf images with four classes: Healthy, Leaf Rust, Phoma, and Cercospora. The research stages included data preprocessing, image augmentation, and dividing the dataset into training, validation, and testing data. The model was trained using the best hyperparameters with 50 epochs, a batch size of 32, the Adam optimizer, and a learning rate of 0.0001. The evaluation results showed that the ResNet-34 model achieved a training accuracy of 99.44%, a validation accuracy of 98.56%, and a testing accuracy of 98.83%. All disease classes were successfully classified with an F1-score above 0.97, with the Phoma class achieving the highest score of 0.9991. The model was then implemented into a mobile application to support real-time coffee leaf disease detection. The results of this study demonstrate that the ResNet-34 architecture is effective in classifying coffee leaf diseases and has the potential for further development as a digital agricultural support system.
Description: 48 Halaman
URI: https://repositori.uma.ac.id/handle/123456789/30262
Appears in Collections:SP - Informatic Engineering

Files in This Item:
File Description SizeFormat 
228160027 - Cessy Cindi Regina Situmorang - Fulltext.pdfCover, Abstract, Chapter I, II, III, V, Bibliography5.13 MBAdobe PDFView/Open
228160027 - Cessy Cindi Regina Situmorang - Chapter IV.pdf
  Restricted Access
Chapter IV489.97 kBAdobe PDFView/Open Request a copy


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