Strawberry Plant Diseases Classification Using CNN Based on MobileNetV3-Large and EfficientNet-B0 Architecture

Pramudhita, Dyah Ajeng and Azzahra, Fatima and Arfat, Ikrar Khaera and Magdalena, Rita and Saidah, Sofia (2023) Strawberry Plant Diseases Classification Using CNN Based on MobileNetV3-Large and EfficientNet-B0 Architecture. Jurnal Ilmiah Teknik Elektro Komputer dan Informatika (JITEKI), 9 (3). pp. 522-534.

Text 2-Strawberry Plant Diseases Classification Using CNN Based on MobileNetV2-Large and EfficientNet-B0 Architecture.pdf Download (1MB)

Abstract

Strawberry is a plant that has many benefits and a high risk of being attacked by pests and diseases. Diseases in strawberry plants can cause a decrease in the quality of fruit production and can even cause crop failure. Therefore, a method is needed to assist farmers in identifying the types of diseases in strawberry plants. Currently, the most popular method for identifying types of disease in strawberry plants automatically is using Convolutional Neural Network (CNN). This study proposed a system to be able to detect strawberry plant diseases by classifying the disease based on leaf images with high accuracy using CNN. The problem with using CNN in the previous studies is the heavyweight architectures that are not suitable for deployment on restricted resource devices. The research contribution is implementing the method with lightweight architectures. The proposed system is a CNN algorithm using MobileNetV3-Large and EfficientNet-B0 models to train pre-processed four-class classification datasets, such as healthy leaves, spider mites pest leaves, caterpillars pest leaves, and powdery mildew leaves. Using those architectures helps the parameters and model size keep in the small condition. The result of this study shows that the MobileNetV3-Large model outperforms EfficientNet-B0. The results obtained the best accuracy reaching 92.14% using the MobileNetV3-Large architecture with the hyperparameter optimizer RMSProp, epochs 70, and learning rate 0.0001. The percentage of the evaluation model using MobileNetV3-Large for precision, recall, and F1-Score achieved 92.81%, 92.14%, and 92.25%. Overall, it presents fairly good results and is felicitous to be deployed on low-power and low-storage devices. Furthermore, in future work, it needs to obtain higher accuracy by generating more datasets with different lighting conditions, trying other augmentation techniques, such as lighting transformation, and proposing a better model.

Item Type:	Artikel Umum
Subjects:	T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisi / Prodi:	Faculty of Industrial Technology (Fakultas Teknologi Industri) > S1-Electrical Engineering (S1-Teknik Elektro)
Depositing User:	M.Eng. Alfian Ma'arif
Date Deposited:	26 Jul 2023 07:08
Last Modified:	26 Jul 2023 07:08
URI:	http://eprints.uad.ac.id/id/eprint/43692

Actions (login required)

View Item