Classification of Downy Mildew Disease in Watermelon Plant Leaf using VGG-16 Convolutional Neural Network

Abstract

Downy Mildew, caused by Pseudoperonospora cubensis, poses a serious threat to watermelon (Citrullus lanatus) crops, with the potential to drastically reduce yields and cause substantial economic losses in the agricultural sector. Early and accurate detection is essential for mitigating these impacts and maintaining crop productivity. An automatic detection system was developed using the VGG-16 Convolutional Neural Network (CNN) model, selected for its high accuracy. Trained and validated on a dataset of 585 images, split into 80% for training and 20% for validation, the model achieved 100% accuracy in both phases, with minimal losses of 0.0325 and 0.0098 respectively. These results underscore the model's potential for precise Downy Mildew detection, supporting improved disease management in precision agriculture.

Plant phenotyping relevance

スイカ葉の病徴を画像からCNNで自動分類する手法が研究の中心であり、植物の病害状態を直接推定するため、植物フェノタイピング手法として採用。

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.