Plant Health Analyzer

Abstract

Timely detection of plant diseases is essential to prevent crop losses and optimize pesticide usage in agriculture. This study proposes an intelligent system, Plant Health Analyzer, for automated plant disease detection using leaf images. The system is based on the EfficientNet-B0 deep learning architecture, known for its high accuracy and computational efficiency. A dataset of 55,448 images from the PlantVillage repository was used for training and evaluation, with appropriate data splitting for validation and testing. The proposed model achieved a validation accuracy of 99.78% and a testing accuracy of 99.76%, demonstrating high reliability in disease classification. A lightweight web-based application was also developed to enable real-time usage, with a model size of only 18 MB, making it suitable for deployment on resource-constrained devices. The results highlight the effectiveness of EfficientNet-B0 for plant disease detection and its potential to support farmers in early diagnosis and decision-making, contributing to advancements in precision agriculture.

Plant phenotyping relevance

葉画像から植物病害状態を推定する深層学習手法を開発・検証し、実利用向けアプリも構築しており、植物フェノタイピング手法が中心である。

Code and data availability

The paper uses the public PlantVillage dataset (55,448 leaf images) and describes an EfficientNet-B0 model and web app, but provides no authors' code, model checkpoint, or dataset deposit URL. The only public asset named (PlantVillage) is a third-party dataset with no URL matching the allowed list, and no availability/

No evidence-backed public reproduction asset is currently recorded.