Scholarlink

<p>As the storage duration of peanuts increases, significant nutritional loss occurs, leading to a gradual decline in quality. The quality changes in peanuts during storage are closely related to the gas information they release. Unsaturated fatty acids and proteins are prone to oxidation and degradation reactions, leading to the release of sour and rancid odors. By analyzing the changes in gas information and combining it with intelligent learning model, an effective technical method can be provided to identify the storage period of peanuts. First, gas data for six peanut samples stored at different periods is collected using the PEN3 electronic nose (e-nose) system under controlled conditions (25&#xa0;°C and 35% RH). Next, a Multi-branch Fusion Attention Mechanism (MBFAM) is developed to highlight key features that influence classification performance of peanut gas data across various storage stages, enhancing the e-nose system’s detection capabilities. Subsequently, MBFAM-Net is designed to optimize classification performance from sensor data acquisition through to the classification process, achieving the highest classification results compared to other models. Ultimately, MBFAM-Net achieves 96.67% accuracy, 97.01% precision, and 96.83% recall. In summary, the combination of the e-nose system and MBFAM-Net offers a reliable method for monitoring and identifying peanut quality.</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

An electronic nose system coupled with an effective deep learning method to identify the different storage periods of peanuts

  • 期刊论文

  • 发表日期：

    2025年4月23日

查看全文
  • Baosheng Wang,
  • Li Zhang

摘要

As the storage duration of peanuts increases, significant nutritional loss occurs, leading to a gradual decline in quality. The quality changes in peanuts during storage are closely related to the gas information they release. Unsaturated fatty acids and proteins are prone to oxidation and degradation reactions, leading to the release of sour and rancid odors. By analyzing the changes in gas information and combining it with intelligent learning model, an effective technical method can be provided to identify the storage period of peanuts. First, gas data for six peanut samples stored at different periods is collected using the PEN3 electronic nose (e-nose) system under controlled conditions (25 °C and 35% RH). Next, a Multi-branch Fusion Attention Mechanism (MBFAM) is developed to highlight key features that influence classification performance of peanut gas data across various storage stages, enhancing the e-nose system’s detection capabilities. Subsequently, MBFAM-Net is designed to optimize classification performance from sensor data acquisition through to the classification process, achieving the highest classification results compared to other models. Ultimately, MBFAM-Net achieves 96.67% accuracy, 97.01% precision, and 96.83% recall. In summary, the combination of the e-nose system and MBFAM-Net offers a reliable method for monitoring and identifying peanut quality.