Scholarlink

<p>To explore a non-destructive identification method for distinguishing Chinese chive (<i>Allium tuberosum</i> Rottl. ex Spreng) seeds from their adulterant, scallion (<i>Allium fistulosum</i> L.) seeds, machine vision and electronic nose technologies were employed. Principal component analysis (PCA), linear discriminant analysis (LDA), artificial neural networks (ANNs), and random forest (RF) algorithms were utilized to perform discriminant analyses based on the acquired data. The comprehensive results indicated that the image-based discrimination method, which integrates PCA with LDA and RF, demonstrated excellent accuracy using the obtained image information. Notably, the RF model established using odor information from the electronic nose achieved the lowest error rates of 0.98% for the training set and 0.70% for the test set. Overall, it was found effective and feasible to apply pattern recognition technology, combining both image and odor information, for the discrimination between Chinese chive seeds and their adulterated scallion seeds.</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

Nondestructive Identification of Chinese Chive Seeds and its Counterfeit Scallion Seeds Based on Machine Vision and Electronic Nose

  • 期刊论文

  • 发表日期：

    2025年1月28日

查看全文
  • Qiang Zhang,
  • Baomei Wu,
  • Weizhong Liu

摘要

To explore a non-destructive identification method for distinguishing Chinese chive (Allium tuberosum Rottl. ex Spreng) seeds from their adulterant, scallion (Allium fistulosum L.) seeds, machine vision and electronic nose technologies were employed. Principal component analysis (PCA), linear discriminant analysis (LDA), artificial neural networks (ANNs), and random forest (RF) algorithms were utilized to perform discriminant analyses based on the acquired data. The comprehensive results indicated that the image-based discrimination method, which integrates PCA with LDA and RF, demonstrated excellent accuracy using the obtained image information. Notably, the RF model established using odor information from the electronic nose achieved the lowest error rates of 0.98% for the training set and 0.70% for the test set. Overall, it was found effective and feasible to apply pattern recognition technology, combining both image and odor information, for the discrimination between Chinese chive seeds and their adulterated scallion seeds.