Scholarlink

To address the economic impact of maize diseases, this work moves beyond time-consuming manual inspection by developing and evaluating automated detection systems. We conduct a comparative study of three distinct deep learning architectures: a custom Deep Convolutional Neural Network (CNN), a fine-tuned VGG16, and an optimized MobileNetV2. Using a dataset of 3852 maize leaf images, our custom CNN achieved a superior testing accuracy of 95.56%, surpassing both the VGG16 (94.30%) and MobileNetV2 (94.00%) models. These results indicate that tailored deep learning models can serve as a highly effective tool for early and precise maize disease identification, aiding in timely agricultural interventions. These findings highlight the effectiveness of deep learning architectures in identifying maize diseases accurately. The proposed models offer a reliable tool for agricultural professionals, enabling timely interventions to safeguard crop health and yield.

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

Maize Plant Disease Detection Using Deep CNN

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Prerit Agarwal,
  • Maaz Alam,
  • Piyush Sahu,
  • Pradip Kumar Sahu,
  • Manas Ranjan Senapati

摘要

To address the economic impact of maize diseases, this work moves beyond time-consuming manual inspection by developing and evaluating automated detection systems. We conduct a comparative study of three distinct deep learning architectures: a custom Deep Convolutional Neural Network (CNN), a fine-tuned VGG16, and an optimized MobileNetV2. Using a dataset of 3852 maize leaf images, our custom CNN achieved a superior testing accuracy of 95.56%, surpassing both the VGG16 (94.30%) and MobileNetV2 (94.00%) models. These results indicate that tailored deep learning models can serve as a highly effective tool for early and precise maize disease identification, aiding in timely agricultural interventions. These findings highlight the effectiveness of deep learning architectures in identifying maize diseases accurately. The proposed models offer a reliable tool for agricultural professionals, enabling timely interventions to safeguard crop health and yield.