Ensemble Deep Learning Classifier for Banana Leaf Disease Detection: A Novel Classifier for Plant Health Monitoring

Abstract

Banana leaf diseases pose significant challenges to agriculture, reducing crop yields and threatening food security. Traditional manual disease detection is often labor-intensive and prone to errors. This study proposes an ensemble deep learning classifier, RF-XG-ResNet50, for automated banana leaf disease identification, combining multiple deep learning architectures to improve accuracy. Data augmentation was employed to enhance training and generalization. The proposed classifier, known as RF- XG-ResNet50 Model, is proposed and compared with two base classifiers like Decision Tree and SVM models, to discriminate between healthy and sick leaves. By automating the disease detection process, this model can potentially reduce human error, save time, and provide realtime insights to farmers, contributing to improved crop management and sustainability. Finally, the proposed classifier gave the highest accuracy of 95% when compared with DT and SVM.

Plant phenotyping relevance

バナナ葉の病徴を画像から分類する深層学習手法を開発・比較しており、植物の病害状態推定が研究の中心であるため。

Code and data availability

公開本文の所在を確認できませんでした。非公開または購読が必要な可能性があります。

No evidence-backed public reproduction asset is currently recorded.