Scholarlink

<p>This study explored the application of a potassium permanganate (KMnO<sub>4</sub>)-based ethylene indicator to assess the freshness of bananas and kiwifruit, alongside the integration of deep learning for interpreting indicator responses. A mobile application was developed using a trained deep learning model for real-time fruit freshness detection. The KMnO<sub>4</sub> concentrations in the indicator were 0.1, 0.5, and 1.0% (w/v), and the fruits were stored at 25&#xa0;&#xa0;C for 10&#xa0;days in polypropylene pouches containing the indicator. The highest correlation was observed between the ethylene concentration and color change in the 0.5% KMnO<sub>4</sub> indicator. The ResNet50 model exhibited high accuracy in predicting the freshness of both fruits. The developed mobile application enabled rapid and accurate real-time detection of freshness by analyzing indicator images of fruit packaging. These results demonstrate that the KMnO<sub>4</sub>-based ethylene indicator is reliable for assessing the freshness of fruits, with mobile software offering a solution for rapid freshness evaluation.</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

Real-time spoilage detection of climacteric fruits using a potassium permanganate-based ethylene indicator and deep learning-based mobile application

  • 期刊论文

  • 发表日期：

    2026年7月12日

查看全文
  • Byeol Yi Kim,
  • Chahn-Mee Moh,
  • Sea Cheol Min

摘要

This study explored the application of a potassium permanganate (KMnO4)-based ethylene indicator to assess the freshness of bananas and kiwifruit, alongside the integration of deep learning for interpreting indicator responses. A mobile application was developed using a trained deep learning model for real-time fruit freshness detection. The KMnO4 concentrations in the indicator were 0.1, 0.5, and 1.0% (w/v), and the fruits were stored at 25  C for 10 days in polypropylene pouches containing the indicator. The highest correlation was observed between the ethylene concentration and color change in the 0.5% KMnO4 indicator. The ResNet50 model exhibited high accuracy in predicting the freshness of both fruits. The developed mobile application enabled rapid and accurate real-time detection of freshness by analyzing indicator images of fruit packaging. These results demonstrate that the KMnO4-based ethylene indicator is reliable for assessing the freshness of fruits, with mobile software offering a solution for rapid freshness evaluation.