Scholarlink

Leaf diseases significantly reduce crop productivity, making early detection essential in precision agriculture. This paper presents a convolutional neural network (CNN)-based approach for automated classification of cotton leaf diseases using an augmented and normalized dataset. The proposed model is trained on a custom dataset of 2,000 images and utilizes image preprocessing, data augmentation techniques, and batch normalization to enhance performance and prevent overfitting. The model achieved a classification accuracy of 99.61%, outperforming existing methods such as Support Vector Machine (SVM) based classifiers and pretrained architectures like VGG-16. Evaluation metrics including precision, recall, and F1-score further validated the model’s robustness. Comparative analysis with recent state-of-the-art methods demonstrates the efficiency and generalizability of the proposed scheme. This study provides a lightweight, high-accuracy solution suitable for real-time deployment in smart farming systems.

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

High-accuracy Cotton Leaf Disease Classification Using a Custom CNN with Data Augmentation and Batch Normalization

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Sambhaji Shivankar,
  • Rushikesh Kolhe,
  • Tushar Mahadik,
  • Krishnaraj Pawar,
  • Vishal Bhosale,
  • Amol Asalekar

摘要

Leaf diseases significantly reduce crop productivity, making early detection essential in precision agriculture. This paper presents a convolutional neural network (CNN)-based approach for automated classification of cotton leaf diseases using an augmented and normalized dataset. The proposed model is trained on a custom dataset of 2,000 images and utilizes image preprocessing, data augmentation techniques, and batch normalization to enhance performance and prevent overfitting. The model achieved a classification accuracy of 99.61%, outperforming existing methods such as Support Vector Machine (SVM) based classifiers and pretrained architectures like VGG-16. Evaluation metrics including precision, recall, and F1-score further validated the model’s robustness. Comparative analysis with recent state-of-the-art methods demonstrates the efficiency and generalizability of the proposed scheme. This study provides a lightweight, high-accuracy solution suitable for real-time deployment in smart farming systems.