Scholarlink

<p>This research&#xa0;investigates the application of deep convolutional neural networks (CNNs) for classifying 5 prominent rice varieties in Bangladesh: Basmati, Chinigura, Jirashail, Kataribhog, and Paijam. Rice is a significant agricultural product in Bangladesh. Proper&#xa0;categorization of rice varieties is vital for quality management, preventing adulteration, and preserving customer confidence. This research leverages advanced CNN architectures (VGG16, VGG19, ResNet50, Xception, DenseNet) and a custom model, BDriceNetworkV0, to automate the classification process. A balanced dataset comprising 15 000 images of rice grains was collected under controlled conditions to train the models. Performance metrics, including&#xa0;accuracy, precision, recall, and F1-score,&#xa0;were used to evaluate the models. VGG19 was identified as the superior performer (accuracy 99.73%). The custom model, BDriceNetworkV0, exhibited potential for resource-constrained applications (accuracy 98.60%); however, with slightly&#xa0;reduced precision. The findings highlight the feasibility of applying deep learning for rice classification, with potential benefits for enhancing quality, reducing operational costs, and improving food security in Bangladesh. This study&#xa0;offers valuable insights into the application of AI in agriculture, particularly in grain classification.</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

Image-based classification of Bangladeshi rice varieties using deep convolutional neural networks

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年1月19日

查看全文
  • Syme Al-Azad Shuvo,
  • Khalilur Rahman,
  • Abdullah Al Naseeh Chowdhury,
  • Gourab Roy,
  • Nafiz Nahid,
  • Mahmud Hasan,
  • Md. Abu Naser Mojumder,
  • Md. Janibul Alam Soeb,
  • Md. Fahad Jubayer

摘要

This research investigates the application of deep convolutional neural networks (CNNs) for classifying 5 prominent rice varieties in Bangladesh: Basmati, Chinigura, Jirashail, Kataribhog, and Paijam. Rice is a significant agricultural product in Bangladesh. Proper categorization of rice varieties is vital for quality management, preventing adulteration, and preserving customer confidence. This research leverages advanced CNN architectures (VGG16, VGG19, ResNet50, Xception, DenseNet) and a custom model, BDriceNetworkV0, to automate the classification process. A balanced dataset comprising 15 000 images of rice grains was collected under controlled conditions to train the models. Performance metrics, including accuracy, precision, recall, and F1-score, were used to evaluate the models. VGG19 was identified as the superior performer (accuracy 99.73%). The custom model, BDriceNetworkV0, exhibited potential for resource-constrained applications (accuracy 98.60%); however, with slightly reduced precision. The findings highlight the feasibility of applying deep learning for rice classification, with potential benefits for enhancing quality, reducing operational costs, and improving food security in Bangladesh. This study offers valuable insights into the application of AI in agriculture, particularly in grain classification.