Scholarlink

Mango belongs to the kingdom Plantae and genus Mangifera, which is commonly known as the “king of fruits”. It is India's most important commercial fruit crop. Due to its large-scale cultivation, mango is vulnerable to several diseases that affect its production. Various microorganisms cause diseases in mangoes at all stages of their growth, starting from seedling stage up to the consumption. These diseases lead to significant yield losses as well as economic losses. Therefore, early detection and effective control measures are essential. This study investigates a transfer learning approach to identify and classify diseases in mango leaves. The Convolutional Neural Network (CNN) models studied include VGG19, NASNetMobile, MobileNetV2, and DenseNet201, applied for the pathogenesis and early identification of seven major diseases that influence the leaves of mango trees, such as Anthracnose, Bacterial Canker, Cutting Weevil, Die Back, Gall Midge, Powdery Mildew, and Sooty Mould. The dataset consists of 4000 images including both healthy and diseased mango leaves. The models were trained and tested on dataset, and confusion matrices were generated for each model. The Highest overall accuracy of 99.02% was achieved by VGG19 model.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Early Identification and Pathogenesis of Leaf Diseases in Magnifera Indica Using Transfer Learning

  • 章节

  • 发表日期：

    2025年

查看全文
  • Arshad Husain,
  • Subhojit Bhowmick

摘要

Mango belongs to the kingdom Plantae and genus Mangifera, which is commonly known as the “king of fruits”. It is India's most important commercial fruit crop. Due to its large-scale cultivation, mango is vulnerable to several diseases that affect its production. Various microorganisms cause diseases in mangoes at all stages of their growth, starting from seedling stage up to the consumption. These diseases lead to significant yield losses as well as economic losses. Therefore, early detection and effective control measures are essential. This study investigates a transfer learning approach to identify and classify diseases in mango leaves. The Convolutional Neural Network (CNN) models studied include VGG19, NASNetMobile, MobileNetV2, and DenseNet201, applied for the pathogenesis and early identification of seven major diseases that influence the leaves of mango trees, such as Anthracnose, Bacterial Canker, Cutting Weevil, Die Back, Gall Midge, Powdery Mildew, and Sooty Mould. The dataset consists of 4000 images including both healthy and diseased mango leaves. The models were trained and tested on dataset, and confusion matrices were generated for each model. The Highest overall accuracy of 99.02% was achieved by VGG19 model.