Scholarlink

<p>One of the most popular tropical fruits, the Mango (Mangifera indica), is famous for its tasty flesh and many health advantages. The king of fruits, mangoes, is in great demand. Therefore, it’s critical to find ways to keep it healthy so we can reap its financial benefits. Because symptoms might vary, automatic leaf disease classification and diagnosis remain difficult. For this reason, we developed a model using deep Learning &amp; Internet of Things to overcome these limitations. The first stage of this model aims at semantic segmentation, which exactly segments the areas of the Alphonso mango leaves affected by certain diseases. The UNet model has residual attention mechanisms in the base; the RAU (Residual Attention UNet) model captures global and local background information, resulting in precise and thorough segmentation results. In the second phase of this model, the segmented regions are inserted into a classification module known as the Attention-MobileNet model. The proposed model’s performance is thoroughly examined using dataset validation, and the outputs show its efficiency in classification and segmentation tasks. We got an accuracy of 98.3%. The model can simultaneously present thorough geometric information about the disease. Here, the disease classification report gives necessary assurance in precision agriculture and thus helps farmers with focused involvement.</p>

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

Attention-based dual-purpose segmentation and classification of diseases in alphonso mango leaves for smart agriculture

  • 期刊论文

  • 发表日期：

    2026年2月20日

查看全文
  • Avinash Pawar,
  • Shankar Deosarkar

摘要

One of the most popular tropical fruits, the Mango (Mangifera indica), is famous for its tasty flesh and many health advantages. The king of fruits, mangoes, is in great demand. Therefore, it’s critical to find ways to keep it healthy so we can reap its financial benefits. Because symptoms might vary, automatic leaf disease classification and diagnosis remain difficult. For this reason, we developed a model using deep Learning & Internet of Things to overcome these limitations. The first stage of this model aims at semantic segmentation, which exactly segments the areas of the Alphonso mango leaves affected by certain diseases. The UNet model has residual attention mechanisms in the base; the RAU (Residual Attention UNet) model captures global and local background information, resulting in precise and thorough segmentation results. In the second phase of this model, the segmented regions are inserted into a classification module known as the Attention-MobileNet model. The proposed model’s performance is thoroughly examined using dataset validation, and the outputs show its efficiency in classification and segmentation tasks. We got an accuracy of 98.3%. The model can simultaneously present thorough geometric information about the disease. Here, the disease classification report gives necessary assurance in precision agriculture and thus helps farmers with focused involvement.