Rapid Classification of Hairtail Fish and Pork Freshness Using an Electronic Nose Based on the PCA Method 英文参考文献.docVIP

Sensors 2012, 12, 260-277; doi:10.3390/s120100260 OPEN ACCESS sensors ISSN 1424-8220 /journal/sensors Article Rapid Classification of Hairtail Fish and Pork Freshness Using an Electronic Nose Based on the PCA Method Xiu-Ying Tian 1,2, Qiang Cai 1,* and Yong-Ming Zhang 2 1 Instrument Lab, Yangtze Delta Region Institute of Tsinghua University, Jiaxing 314006, Zhejiang, China; E-Mail: xiuying.tian@ 2 Department of Life and Environmental Science, Shanghai Normal University, Shanghai 200237, China; E-Mail: zhym@ * Author to whom correspondence should be addressed; E-Mail: caiq@; Tel.: +86-573-8258-6608; Fax: +86-573-8258-6611. Received: 22 November 2011; in revised form: 18 December 2011 / Accepted: 26 December 2011 / Published: 28 December 2011 Abstract: We report a method for building a simple and reproducible electronic nose based on commercially available metal oxide sensors (MOS) to monitor the freshness of hairtail fish and pork stored at 15, 10, and 5 °C. After assembly in the laboratory, the proposed product was tested by a manufacturer. Sample delivery was based on the dynamic headspace method, and two features were extracted from the transient response of each sensor using an unsupervised principal component analysis (PCA) method. The compensation method and pattern recognition based on PCA are discussed in the current paper. PCA compensation can be used for all storage temperatures, however, pattern recognition differs according to storage conditions. Total volatile basic nitrogen (TVBN) and aerobic bacterial counts of the samples were measured simultaneously with the standard indicators of hairtail fish and pork freshness. The PCA models based on TVBN and aerobic bacterial counts were used to classify hairtail fish samples as “fresh” (TVBN ≤ 25 g and microbial counts ≤ 106 cfu/g) or “spoiled” (TVBN ≥ 25 g and microbial counts ≥ 106 cfu/g) and