Scholarlink

<p>Fresh fruits and vegetables are crucial for human health, supplying diverse nutrients. Fruits have various quality attributes, including texture, taste, and external and internal characteristics, yet they are prone to damage and decay. Key chemical indicators used to judge the fruit quality include soluble solid content (SSC) and pH. SSC is essential for determining the ripeness and quality of fruits, while pH measures acidity, with lower values indicating higher acid content. With advancements in sensor technology, non-destructive methods for assessing fruit quality have become more prevalent. Hyperspectral imaging is now widely used for detecting soluble solids and pH in fruits, providing comprehensive spectral data. This paper thoroughly reviews the integration of HSI and deep learning techniques in fruit quality analysis, focusing on non-destructive methods for assessing chemical and physical parameters such as SSC, pH, and firmness. Hyperspectral imaging, with its detailed spectral data, offers new possibilities for precise and automated quality assessment. Deep learning algorithms, particularly convolutional neural networks, have shown promise in enhancing the accuracy of fruit quality prediction models. The review covers various studies employing DL-based architectures like 1D/3D ResNet, SAE-RF, and PCA-CNN across a wide range of fruits, highlighting the improvements in performance over traditional methods. Challenges, including the high dimensionality of HSI data, hardware costs, and scalability, are discussed, as well as potential future trends in the field. This manuscript concludes with recommendations for advancing HSI-based DL applications for more reliable, scalable, and real-time fruit quality assessment systems.</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

Recent trends in deep learning and hyperspectral imaging for fruit quality analysis: an overview

  • 期刊论文

  • 发表日期：

    2025年8月16日

查看全文
  • Neha P. Lanke,
  • Manoj B. Chandak

摘要

Fresh fruits and vegetables are crucial for human health, supplying diverse nutrients. Fruits have various quality attributes, including texture, taste, and external and internal characteristics, yet they are prone to damage and decay. Key chemical indicators used to judge the fruit quality include soluble solid content (SSC) and pH. SSC is essential for determining the ripeness and quality of fruits, while pH measures acidity, with lower values indicating higher acid content. With advancements in sensor technology, non-destructive methods for assessing fruit quality have become more prevalent. Hyperspectral imaging is now widely used for detecting soluble solids and pH in fruits, providing comprehensive spectral data. This paper thoroughly reviews the integration of HSI and deep learning techniques in fruit quality analysis, focusing on non-destructive methods for assessing chemical and physical parameters such as SSC, pH, and firmness. Hyperspectral imaging, with its detailed spectral data, offers new possibilities for precise and automated quality assessment. Deep learning algorithms, particularly convolutional neural networks, have shown promise in enhancing the accuracy of fruit quality prediction models. The review covers various studies employing DL-based architectures like 1D/3D ResNet, SAE-RF, and PCA-CNN across a wide range of fruits, highlighting the improvements in performance over traditional methods. Challenges, including the high dimensionality of HSI data, hardware costs, and scalability, are discussed, as well as potential future trends in the field. This manuscript concludes with recommendations for advancing HSI-based DL applications for more reliable, scalable, and real-time fruit quality assessment systems.