Enhanced deep learning model architecture for plant disease detection in Chilli plants

Abstract

A new deep-learning model for classifying and detecting plant diseases in chilli plants is described. It is built on a modified version of the MobileNet architecture. The model overcomes conventional diagnostic tools’ high computing costs and restricted adaptability by combining sophisticated optimisation models and reliable training procedures. The model considerably reduces the time and resources needed for an accurate diagnosis while effectively managing complicated illness presentations, with a diagnostic accuracy of 97.18%. Using the chilli leaf picture dataset, data augmentation, and finetuning techniques, the model shows promise for real-time disease diagnosis in agricultural environments. The study underscores the importance of high-quality image data and extensive training datasets, calling for further evaluation across various climatic and environmental conditions to ensure robustness and adaptability. This research opens new opportunities for AI-based models in diverse agricultural contexts, potentially leading to significant advancements in precision farming.

Plant phenotyping relevance

唐辛子葉画像から植物病害を検出・分類する深層学習モデルの開発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に該当する。

Code and data availability

The paper describes a chilli leaf image dataset (300 photos, 5 classes) and a modified MobileNet model, but provides no public URL, repository, or availability statement for the dataset, code, or trained model. All URLs in the text are citations to prior work or author profile pages, none of which host paper-specific.

No evidence-backed public reproduction asset is currently recorded.