Improved Agricultural Monitoring through Automated Detection and Classification of Purple Blotch and Yellow Blotch Diseases on Onion Leaves utilizing Dual Branch Graph Neural Network

Abstract

Traditional visual inspection is replaced by plant disease detection (PDD), which integrates plant pathology, image processing, remote sensing, agronomy and machine learning. Remote sensing tools and high-resolution drone imagery enhance real-time disease monitoring. However, challenges like data scarcity and algorithm generalization make accurate and automated disease diagnosis more complex. In this manuscript, Improved Agricultural Monitoring through Automated Detection and Classification of Purple Blotch and Yellow Blotch Diseases on Onion Leaves utilizing Dual Branch Graph Neural Network (ADC-PBYBD-OL-DBGNN) is proposed. The onion leaf dataset was first collected in the Karnataka village of Chilwadigi, which has a variety of climates. Then the collected images are preprocessed using Distributed Adaptive Spatial Filtering (DASF) is used for resizing and cropping. The preprocessed images are fed to Onion Leaf Disease Detection utilizing Dual Branch Graph Neural Network (DBGNN) for detecting and classifying as Healthy, Iris Yellow Virus, Purple Blotch and Leaf Blight. DBGNN generally doesn't show any adaptation of optimization techniques for figuring out the best parameters to guarantee precise onion leaf detection. Hence, Superb Fairy-wren Optimization Algorithm (SFOA) is utilized to optimize DGRGNN for precisely classifying the onion leaves. Then the proposed ADC-PBYBD-OL-DBGNN is implemented and the performance metrics like Accuracy, Precision, Recall, F1 score, Computational time are analyzed. Finally, the performance of proposed ADC-PBYBD-OL-DBGNN method provides 26.68%, 25.75%, and 26.16% higher accuracy and 27.49%, 24.75%, and 25.85% higher precision while compared with existing methods such as the onion plant leaf image dataset for classification and detection (OPLM-CD-Xpection), the onion and maize image datasets for creating AI-based classification models for pests and diseases (OMI-PD-CNN) and the meta-learning based neural architecture search for automated generation of deep neural networks for plant disease detection tasks (SAG-PDD-DNN) correspondingly.

Plant phenotyping relevance

タマネギ葉の画像から病害状態を検出・分類する画像ベースの植物フェノタイピング手法を提案し、性能比較まで行っており、方法開発が中心である。

Code and data availability

植物フェノタイピング解析を再現する公開資産であることを、入力本文と直接リンクから確認できなかったため保留しました。

No evidence-backed public reproduction asset is currently recorded.