Klasifikasi Penyakit Tanaman Kopi Robusta Berdasarkan Citra Daun Menggunakan Transfer Learning Dan Convolutional Neural Network

Zahra, Fathya (2024) Klasifikasi Penyakit Tanaman Kopi Robusta Berdasarkan Citra Daun Menggunakan Transfer Learning Dan Convolutional Neural Network. Diploma thesis, Institut Teknologi Sepuluh Nopember.

Text 5003201164-SARJANA_Skripsi.pdf - Accepted Version Restricted to Repository staff only until 1 October 2026. Download (4MB) | Request a copy

Abstract

Negara penghasil kopi terbesar ketiga di dunia adalah Indonesia. Produksi kopi memegang peran yang sangat penting, terutama untuk jenis biji kopi Robusta yang paling banyak diproduksi di negara Indonesia. Namun, terdapat tantangan dalam produksi kopi Robusta di Indonesia, seperti rendahnya produktivitas dan mutu yang tidak memenuhi standar ekspor akibat serangan hama, jamur, dan bakteri. Salah satu cara paling efektif untuk mengidentifikasi penyakit pada tanaman kopi Robusta adalah dengan mengamati kondisi permukaan daun. Oleh karena itu, penelitian ini akan melakukan klasifikasi citra daun kopi Robusta menggunakan dua metode, yaitu Convolutional Neural Network (CNN) dan Transfer Learning dengan arsitektur Inception V3 dan Visual Geometry Group (VGG16). Data pelatihan akan diterapkan pada ketiga model tersebut, dan masing-masing model akan dilatih dengan kombinasi hyperparameter yang berbeda untuk menemukan performa terbaik dari setiap model. Setelah perbandingan dilakukan, model terbaik ditemukan, yaitu Inception V3, dengan kombinasi hyperparameter yang meliputi learning rate sebesar 0,0001, optimizer Adam, dropout 0,2, dan batch size 64. Model dengan kombinasi tersebut mencapai akurasi data pelatihan sebesar 0,9748, nilai loss data pelatihan sebesar 0,1101, akurasi data validasi sebesar 0,9587, dan nilai loss pada data validasi sebesar 0,0517. Model tersebut menunjukkan performance yang baik dalam mengklasifikasikan objek citra daun pada tanaman kopi Robusta.
========================================================================================================================
Indonesia is the third-largest coffee-producing country in the world. Coffee production plays a very important role, especially for Robusta coffee beans, which are the most widely produced in Indonesia. However, there are challenges in Robusta coffee production in Indonesia, such as low productivity and quality that does not meet export standards due to attacks by pests, fungi, and bacteria. One of the most effective ways to identify diseases in Robusta coffee plants is by observing the condition of the leaf surface. Therefore, this research will classify images of Robusta coffee leaves using two methods: Convolutional Neural Network (CNN) and Transfer Learning with the architectures of Inception V3 and Visual Geometry Group (VGG16). Training data will be applied to these three models, and each model will be trained with different combinations of hyperparameters to find the best performance for each model. After comparison, the best model was found to be Inception V3, with a combination of hyperparameters including a learning rate of 0.0001, Adam optimizer, dropout 0.2, and batch size 64. This model achieved a training data accuracy of 0.9748, training data loss of 0.1101, validation data accuracy of 0.9587, and validation data loss of 0.0517 This model shows good performance in classifying leaf images of Robusta coffee plants.

Item Type:	Thesis (Diploma)
Uncontrolled Keywords:	CNN, klasifikasi, Kopi Robusta, Transfer learning, classification, Robusta Coffee.
Subjects:	Q Science Q Science > Q Science (General) > Q325.5 Machine learning. Support vector machines. Q Science > Q Science (General) > Q325.78 Back propagation
Divisions:	Faculty of Science and Data Analytics (SCIENTICS) > Statistics > 49201-(S1) Undergraduate Thesis
Depositing User:	Fathya Zahra
Date Deposited:	09 Aug 2024 05:26
Last Modified:	09 Aug 2024 05:26
URI:	http://repository.its.ac.id/id/eprint/114983

Actions (login required)

View Item