Prediction of barberry witches' broom rust disease using artificial intelligence models: a case study in South Khorasan, Iran.

Abstract

The South Khorasan Province in Iran is the main producer of seedless barberry, accounting for 98% of the country's production. This has led to significant economic growth in the region. However, the cultivation of barberry is threatened by the rust fungus Puccinia arrhenatheri, which causes witches' brooms on Berberis vulgaris L. var. asperma. Our research aims to detect infected leaves containing this fungal pathogen using deep learning (DL)-based artificial intelligence (AI) techniques on an available dataset. We captured healthy and infected barberry foliage images and used conventional laboratory methods to label them. We developed a convolutional neural network (CNN) deep learning model using TensorFlow's Keras API to detect and classify barberry broom rust disease. A cross-validation technique is used to check the robustness of the proposed model. The results imply that the proposed model successfully distinguished between healthy specimens and those affected by broom rust disease. The model achieved an impressive accuracy rate of 98% in automatically identifying the disease type and its severity. This interdisciplinary research demonstrates the practical application of AI in agriculture, providing timely intervention strategies to protect crop yields and maintain economic viability in the face of plant diseases.

Plant phenotyping relevance

植物葉の画像から感染状態と病害の種類・重症度を推定するCNN手法を開発し、交差検証で頑健性を評価しており、病害表現型の取得・推定が中心である。

Code and data availability

The paper's barberry leaf image dataset (208 healthy, 549 broom rust images) and CNN analysis code are not publicly deposited; the Data availability statement says they are available from the corresponding author upon reasonable request.

No evidence-backed public reproduction asset is currently recorded.