Optimized sequential model for superior classification of plant disease.

Abstract

Indian agriculture is vital sector in the country's economy, providing employment and sustenance to millions of farmers. However, Plant diseases are a serious risk to crop yields and farmers' livelihoods. Traditional plant disease diagnosis methods rely heavily on human expertise, which can lead to inaccuracies due to the invisible nature of early disease symptoms and the labor-intensive process, making them inefficient for large-scale agricultural management. To recover from this and, address these challenges, this study explores deep learning, specifically Convolutional Neural Networks (CNN), as a means to enhance the accuracy and efficiency of plant disease detection. Deep learning architectures, like convolutional neural network, can autonomously learn and extract complicated characteristics and patterns from huge datasets. Our research, conducted on mango and groundnut leaves collected during field visits in western Maharashtra and supplemented by online datasets, demonstrates a CNN model that achieves an impressive 96% accuracy as compared to machine learning techniques that follow tedious feature extraction. Furthermore, image processing contributes to enhancing the dataset through normalization, resizing, and augmentation for better classification results. Overall, CNN can continuously improve and adapt its performance through iterative training, resulting in higher accuracy rates and reduced false positives in contrast to conventional machine learning methods.

Plant phenotyping relevance

CNNによる植物葉の病害状態分類と画像前処理・性能比較が研究の中心であり、植物病害表現型の取得・推定手法に該当する。

Code and data availability

The paper's field-collected mango/groundnut leaf image dataset (supplemented by Mendeley and Kaggle) and analysis code are not publicly deposited; the Data availability statement says datasets are available only from the corresponding author on reasonable request. No public URL, repository, or identifier for a paper-­‐

No evidence-backed public reproduction asset is currently recorded.