Klasifikasi Penyakit Daun Padi Menggunakan Support Vector Machine Berdasarkan Fitur Mendalam (Deep Feature)

Margarita, Devina (2024) Klasifikasi Penyakit Daun Padi Menggunakan Support Vector Machine Berdasarkan Fitur Mendalam (Deep Feature). Undergraduate thesis, UPN Veteran Jawa Timur.

	Text (Cover) 20081010009-Cover.pdf Download (780kB)
	Text (Bab 1) 20081010009_Bab 1.pdf Download (260kB)
	Text (Bab 2) 20081010009-Bab 2.pdf Restricted to Repository staff only until 17 July 2026. Download (1MB)
	Text (Bab 3) 20081010009-Bab 3.pdf Restricted to Repository staff only until 17 July 2026. Download (2MB)
	Text (Bab 4) 20081010009-Bab 4.pdf Restricted to Repository staff only until 17 July 2026. Download (2MB)
	Text (Bab 5) 20081010009-Bab 5.pdf Download (229kB)
	Text (Daftar Pustaka) 20081010009-Daftar Pustaka.pdf Download (238kB)
	Text (Lampiran) 20081010009-Lampiran.pdf Restricted to Repository staff only Download (518kB)

Abstract

Rice as a staple food in Indonesia, especially rice, is experiencing a decline production up to 53.63 million tons of GKG in 2023, a decrease of around 2.05% from previous year. According to FAO, 20-40% of world food production fails caused by pests and diseases. Detection of pest attacks on rice plants can be done. This is done by observing the condition of the leaves directly. However, this is vulnerable against mistakes, especially among farmers who are old and his vision decreased. Technologies such as image processing, machine learning, and deep learning can used to identify diseases in rice leaves. This research using the CNN method with AlexNet and VGG19 architecture for extraction image features, as well as classification using SVM with the SMO optimization algorithm (Sequential Minimal Optimization). CNN is effective in recognizing visual patterns of images and not limited to focusing on one feature extraction pattern, so it is suitable for various types image. AlexNet has a relatively simple structure but is still capable gives good results in feature extraction, while VGG19 has more many layers compared to AlexNet. By using these two architectures, can be compared which is more effective in reading data in detail or more generally. The SVM method with SMO optimization can divide the problem large optimization into a series of smaller sub-problems, resulting in a process training and testing can be faster and more efficient which allows for handles datasets with a larger number of samples and features with more effective. By using this combination, it is possible to detect pest attacks done more accurately. The results show an accuracy of 98.49% using deep feature extraction from AlexNet and the SVM setup use a polynomial kernel with values hyperparameters C=100, gamma=0.01, degree=1, and coef=0, with total computation time being 68.49 seconds. Use of AlexNet architecture with settings Appropriate hyperparameters in SVM enable the model to recognize patterns image of rice leaves well, resulting in high accuracy. The simplicity of AlexNet makes it more efficient in terms of computing, reducing the time required for training and testing, and also obtain good accuracy. Keywords: Image of rice leaf disease, Convolutional Neural Network, Support Vector Machine, Sequential Minimal Optimization, AlexNet, VGG19

Item Type:

Thesis (Undergraduate)

Contributors:

Contribution	Contributors	NIDN/NIDK	Email
Thesis advisor	Maulana, Hendra	NIDN1423128301	hendra.maulana.if@upnjatim.ac.id
Thesis advisor	Mandyartha, Eka Prakarsa	NIDN0725058805	eka_prakarsa.fik@upnjatim.ac.id

Subjects:

Q Science > QA Mathematics > QA76.6 Computer Programming
Q Science > QA Mathematics > QA76.87 Neural computers

Divisions:

Faculty of Computer Science > Departemen of Informatics

Depositing User:

Devina Devina Margarita

Date Deposited:

17 Jul 2024 06:35

Last Modified:

17 Jul 2024 06:35

URI:

https://repository.upnjatim.ac.id/id/eprint/26308