Detection of total viable count in spiced beef using hyperspectral imaging combined with wavelet transform and multiway partial least squares algorithm. Issue 1 (23rd August 2017)

• Record Type: Journal Article
• Authors: Yang, Dong; Lu, Anxiang; Ren, Dong; Wang, Jihua
• Is Part Of: Journal of food safety. Volume 38:Issue 1(2018)
• Page Start: n/a
• Page End: n/a
• Publication Date: 2017-08-23
• Subjects: Food adulteration and inspection -- Periodicals; Food contamination -- Periodicals; Food -- Analysis -- Periodicals; Food -- Microbiology -- Periodicals; Pathogenic bacteria -- Periodicals; Food handling -- Periodicals; Food preservatives -- Periodicals; 664
• DOI: 10.1111/jfs.12390
• Languages: English
• ISSNs: 0149-6085

Abstract

Abstract: The feasibility of using hyperspectral imaging technology combined with wavelet transform and multiway partial least squares (N‐PLS) algorithm to predict the total viable count (TVC) of spiced beef during storage was investigated. The mean spectral data were extracted from the hyperspectral images and further decomposed in nine levels by daubechies8 (db8) wavelet function to obtain an approximation coefficient (A9) and nine detail coefficients (D1–D9). Selecting wavelet coefficients to compose different three‐dimension matrixes, further integrating N‐PLS algorithm to establish the predictive models for detecting TVC in spiced beef. The experimental results show that the N‐PLS model with D4, D5, D6, D7 coefficient (also named D4, 5, 6, 7 ‐N‐PLS) exhibit an excellent prediction capability for TVC of spiced beef sample with a higher determination coefficients in prediction ( R p 2 ) of 0.934 and lower root mean squared errors estimated by prediction of 0.755 than other N‐PLS models, raw spectra PLS model, and Unfold‐PLS models. Therefore, the established model using three‐dimension data array and N‐PLS algorithm has great potential in the TVC value detection of spiced beef and other meat productions. Practical applications: Spiced beef is one of the most popular meat products in China owing to its good taste, low fat, and high protein. Unfortunately, the spiced beef during storage is just a suitable habitat for many pathogens and spoilage microbes to colonize. Thus, it is necessary to monitor the microbiological contamination to guarantee the sanitary quality of meat productions. The total viable count (TVC) of bacteria was considered as a key indicator for the freshness evaluation of meat productions, Traditional techniques for determination of TVC are laborious, time consuming, and unsuitable for modern meat industrial processing and production technologies. Therefore, this study attempted to use hyperspectral imaging technology combined with wavelet transform and multiway partial least squares (N‐PLS) algorithm to predict the TVC of spiced beef during storage. The proposed method is helpful for most consumers to judge easily the freshness state of meat productions and further purchase them.