Towards Automated Detection of Tomato Leaf Diseases

Prompt detection of plant diseases is crucial due to their detrimental impact on society, ecology, and economy. Accurate categorization of tomato leaf diseases is essential for sustaining agricultural productivity and guaranteeing food security. Consequently, it is exceedingly intricate and leads to...

Full description

Saved in:

Bibliographic Details
Published in:	2024 6th International Conference on Electrical Engineering and Information & Communication Technology (ICEEICT) pp. 387 - 392
Main Authors:	Ahmed, Md. Redwan, Haque, Rezaul, Rahman, S M Arafat, Afridi, Sheikh, Abir, Md Fahim Faez, Hossain, MD Forhad, Khan, M. Saddam Hossain, Uddin, Md. Mohsin
Format:	Conference Proceeding
Language:	English
Published:	IEEE 02-05-2024
Subjects:	agricultural environments android application Biological system modeling deep learning Economics Food security plant disease Plant diseases Predictive models Productivity Transfer learning
Online Access:	Get full text
Tags:	Add Tag No Tags, Be the first to tag this record!

Description
Summary:	Prompt detection of plant diseases is crucial due to their detrimental impact on society, ecology, and economy. Accurate categorization of tomato leaf diseases is essential for sustaining agricultural productivity and guaranteeing food security. Consequently, it is exceedingly intricate and leads to an overwhelming amount of labor and waste of time. This paper utilized deep learning with novel ensemble architectures to categorize nine diseases affecting tomato plant leaves, as well as healthy leaves. This method utilized a grand total of 60,000 photos. The paper introduces a new method for classifying tomato leaf diseases by utilizing a hybrid transfer learning model. The approach we employ combines the advantages of transfer learning with conventional machine learning methods to enhance the accuracy and efficiency of classification. We assess our methodology by utilizing a dataset consisting of 60,000 photos of tomato leaves afflicted with ten distinct illnesses. The experimental findings show that our XSVC model achieves a remarkable accuracy of 99.51%, surpassing the performance of the most advanced approaches available. In addition, we provide an Android application that enables users to upload photos of tomato leaves for disease prediction utilizing the XSVC model. Our research enhances the field of plant disease diagnosis by providing precise and effective solutions for promptly detecting and categorizing diseases in agricultural environments.
ISSN:	2769-5700
DOI:	10.1109/ICEEICT62016.2024.10534559