Xu Weidong, He Yingchao, Li Jiaheng, Deng Yong, Zhou Jianwei, Xu Enbo, Ding Tian, Wang Wenjun, Liu Donghong
College of Biosystems Engineering and Food Science, National-Local Joint Engineering Laboratory of Intelligent Food Technology and Equipment, Zhejiang Key Laboratory for Agro-Food Processing, Zhejiang Engineering Laboratory of Food Technology and Equipment, Fuli Institute of Food Science, Zhejiang University, Hangzhou 310058, China.
Ningbo Research Institute, Zhejiang University, Ningbo 315100, China; Zhejiang University Ningbo Institute of Technology, Ningbo 315100, China.
Meat Sci. 2022 Dec;194:108950. doi: 10.1016/j.meatsci.2022.108950. Epub 2022 Aug 19.
Beef is easily spoiled, resulting in foodborne illness and high societal costs. This study proposed a novel olfactory visualization system based on colorimetric sensor array and chemometric methods to detect beef freshness. First, twelve color-sensitive materials were immobilized on a hydrophobic platform to acquire scent information of beef samples according to solvatochromic effects. Second, machine vision algorithms were used to extract the scent fingerprints, and principal component analysis (PCA) was employed to compress the feature dimensions of the fingerprints. Finally, four qualitative models, k-nearest neighbor, extreme learning machine, support vector machine (SVM), and random forest, were constructed to evaluate the beef freshness according to the value of total volatile basic nitrogen (TVB-N) and total viable counts (TVC). Results demonstrated that SVM had a preferable prediction ability, with 95.83% and 95.00% precision in the training and prediction sets, respectively. The results revealed that the simple constructed olfactory visualization sensor system could rapidly, robustly, and accurately assess beef freshness.
牛肉容易变质，会导致食源性疾病并造成高昂的社会成本。本研究提出了一种基于比色传感器阵列和化学计量学方法的新型嗅觉可视化系统，用于检测牛肉的新鲜度。首先，将十二种颜色敏感材料固定在疏水平台上，根据溶剂化显色效应获取牛肉样品的气味信息。其次，使用机器视觉算法提取气味指纹，并采用主成分分析（PCA）来压缩指纹的特征维度。最后，构建了四种定性模型，即k近邻、极限学习机、支持向量机（SVM）和随机森林，根据总挥发性盐基氮（TVB-N）值和总活菌数（TVC）来评估牛肉的新鲜度。结果表明，SVM具有较好的预测能力，在训练集和预测集中的准确率分别为95.83%和95.00%。结果表明，所构建的简单嗅觉可视化传感器系统能够快速、稳健且准确地评估牛肉的新鲜度。
