Scholarlink

Agricultural pests cause significant damage to food crops globally, leading to food scarcity for the human population. Timely identification of pests is crucial for selecting appropriate pest control methods, which is not possible manually. Deep Learning methods have proved very efficient in this regard, and many pre-trained Convolutional Neural Network models, like ResNet, DenseNet, VGGNet, and Inception-based models, are available for image processing tasks. This paper proposes a statistical approach and oversampling using geometric transformations to reduce data imbalance in the IP102 dataset and implement multiple pre-trained CNN models to classify pest species to improve prediction accuracy.

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

Agricultural Pest Classification Using Dense Network with Statistical Approach to Dataset Balancing

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Neetu Agrawal,
  • Mehul Mahrishi,
  • Mukesh Kumar Gupta

摘要

Agricultural pests cause significant damage to food crops globally, leading to food scarcity for the human population. Timely identification of pests is crucial for selecting appropriate pest control methods, which is not possible manually. Deep Learning methods have proved very efficient in this regard, and many pre-trained Convolutional Neural Network models, like ResNet, DenseNet, VGGNet, and Inception-based models, are available for image processing tasks. This paper proposes a statistical approach and oversampling using geometric transformations to reduce data imbalance in the IP102 dataset and implement multiple pre-trained CNN models to classify pest species to improve prediction accuracy.