Scholarlink

<p>Cotton serves as a crucial economic crop and raw material for the global textile industry. However, cotton leaf diseases such as brown spot, verticillium wilt, target spot, and fusarium wilt impair cotton yield and quality. For accurate early detection of cotton leaf diseases, we propose an integrated deep learning framework that integrates an improved ResNeXt convolutional neural network (CNN) with the efficient channel attention (ECA) module and call it RECA. Specifically, we first extracted fundamental and critical feature information using two convolutional layers paired with max-pooling layers. Second, we modified the ResNeXt network to reduce the number of computational parameters. Third, we incorporated three ECA modules into ResNeXt to enhance classification accuracy. Finally, the Softmax classifier was utilized to determine the categories of cotton leaf diseases. Experimental results show that the proposed method achieves an overall classification accuracy of 99.04%, outperforming Efficient-B7, Inception-V4, and ShuffleNet-V2. The developed system enables timely and accurate diagnosis of plant leaf diseases, offering valuable references for agricultural production and plant pathological analysis. Additionally, this strategy provides insights for adapting deep learning models to mobile on-field plant disease detection devices.</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

RECA: an efficient deep learning model for the rapid identification of cotton leaf diseases

  • 期刊论文

  • 发表日期：

    2026年3月9日

查看全文
  • Jingfang Wang,
  • Hongyun Song,
  • Madineh Bijani,
  • Longguo Wu,
  • Laixiang Xu

摘要

Cotton serves as a crucial economic crop and raw material for the global textile industry. However, cotton leaf diseases such as brown spot, verticillium wilt, target spot, and fusarium wilt impair cotton yield and quality. For accurate early detection of cotton leaf diseases, we propose an integrated deep learning framework that integrates an improved ResNeXt convolutional neural network (CNN) with the efficient channel attention (ECA) module and call it RECA. Specifically, we first extracted fundamental and critical feature information using two convolutional layers paired with max-pooling layers. Second, we modified the ResNeXt network to reduce the number of computational parameters. Third, we incorporated three ECA modules into ResNeXt to enhance classification accuracy. Finally, the Softmax classifier was utilized to determine the categories of cotton leaf diseases. Experimental results show that the proposed method achieves an overall classification accuracy of 99.04%, outperforming Efficient-B7, Inception-V4, and ShuffleNet-V2. The developed system enables timely and accurate diagnosis of plant leaf diseases, offering valuable references for agricultural production and plant pathological analysis. Additionally, this strategy provides insights for adapting deep learning models to mobile on-field plant disease detection devices.