Predicting liquid loss of frozen and thawed cod from hyperspectral imaging

Abstract

As the ability to appraise the quality of every fish in a delivery in a consistent, objective, and rapid manner has numerous advantages for both sellers and buyers, there has been much research into methods to achieve this. One possible proxy for quality assessment is liquid loss, which correlates with undesirable sensory attributes. This study evaluated whether hyperspectral imaging could predict liquid loss on samples that had undergone a program of freezing and thawing. Vacuum-packaged cod loins were split into two groups, one which was kept chilled and the other that underwent a program of freezing and thawing. Multivariate analysis of the hyperspectral imaging data on the chilled samples could predict liquid loss with good accuracy for samples that underwent no further processing. Analysis of data from the hyperspectral images of the frozen and thawed cod samples also showed a good ability to predict their liquid loss. These results indicate that hyperspectral imaging is a promising method for non-invasive quality monitoring of cod products in different processing states. Also, whereas previous research had been unable to predict sample processing protocols, improvements to the hyperspectral imaging technology now enables identification of samples based on freezing and thawing procedures.

Recent advances in shelf life prediction models for monitoring food quality

“…Coconut water (Srivastava et al, 2019) Zero-order reaction First-order reaction RBF neural network Shi et al (2019Shi et al ( , 2018 developed two RBF models to predict the freshness of stored tilapia fillets under nonisothermal conditions using the optimal wavelength of HSI and an integration of electronic nose and tongue to validate the ability of model to predict the freshness index of stored fillets. For other fish, Anderssen et al (2020) used RBF combined with HSI for the predictive analysis of liquid loss in vacuum-packaged cod fillets in chilled, frozen, and thawed states in different processing stages. Kong et al (2016) used RBF to model the mass variation of brined common carp, determined the study index based on changes such as lipid hydrolysis, and repeatedly tested the best model to successfully predict the shelf life with a 5% error.…”

“…For other fish, Anderssen et al. (2020) used RBF combined with HSI for the predictive analysis of liquid loss in vacuum‐packaged cod fillets in chilled, frozen, and thawed states in different processing stages. Kong et al.…”

Study on the Diffusion and Optimization of Sucrose in Gaido Seak Based on Finite Element Analysis and Hyperspectral Imaging Technology

“…Hyperspectral images contain hundreds of continuous bands of images, which contain not only the spectral information of the sample, but also a lot of useless background spectral information, making the spectral images large in data and redundant in information [14]. After acquiring the spectral information of a sample, different pre-processing methods are used to correct the full spectrum in time to eliminate scattering effects and useless high-frequency noise interference during the shooting process and to highlight the characteristic peaks in the spectral information [22]. Five different pre-processing methods were used in this study including standard normal variate information (SNVT), convolutional smoothing (Savitzky-golay, SG), multiplicative scatter correction (MSC), first derivative (1st-Der) and second derivative (2nd-Der) as shown in Figure S2.…”

Quality and Shelf-Life Modeling of Frozen Fish at Constant and Variable Temperature Conditions

“…Since storage temperatures in the real cold chain vary, it is necessary to be able to estimate reliably the effect of a known time-temperature scenario that may occur in the real frozen food supply chain on the shelf life of the tested frozen food [ 7 , 8 ]. Despite the many years of research, only a few methods have been employed for determining of quality attributes of frozen fish [ 2 , 9 , 10 , 11 , 12 , 13 , 14 , 15 , 16 , 17 ]. The systematic monitoring and modeling of the temperature dependence would be an essential prerequisite for shelf-life optimization and effective management of the cold chain [ 4 , 18 , 19 ], using conventional temperature control and monitoring and intelligent packaging applications such as TTIs (Time Temperature Integrators) [ 20 , 21 , 22 ].…”