Monitoring of fishery product quality using an electronic nose and visible/near-infrared spectroscopy.

摘要

In order to evaluate new technologies that could improve quality determination of fishery products, this research investigated the application of electronic noses (e-nose) and Visible/Near Infrared (VIS/NIR) spectroscopy as possible sensing technologies. The quality of fishery products has always been hard to define, and is typically based on the general perception of the consumer evaluating the product. Expiration dates serve as a guide, but the sensory appeal of a fishery product is generally the deciding factor as to whether a product is deemed acceptable or not by the end consumer.; Various chemical and sensory methods to determine fish freshness are available to the food industry, but most are expensive, time consuming or destructive. A rapid, nondestructive method to ascertain fish quality would be of great benefit to both the industry that is eager to provide its consumers with a fresh, safe product and the consumer who is increasingly looking for a better guarantee of food quality.; The multivariate analysis (MVA) techniques used by both e-nose and VIS/NIR technologies are similar, but widespread use of these techniques has only become possible with the increased computing power of the past few years.; E-nose technology is a slightly newer technology than VIS/NIR. Due to its more recent introduction, a rapid decay study was first performed to evaluate the feasibility of this method to sense decay time in a readily available fishery product of tilapia. Linear discriminate analysis (LDA) was used as a feature extraction method. This allows for the class of a given sample to be taken into account when features are extracted to yield a much better model. Separating the samples into 6 hour classes, a classification rate of 97.8% was achieved.; Once the e-nose had shown promise in quantifying fish decay, a more continuous model was chosen to more accurately model the continuous decay of fish products. It was also decided to perform the testing at actual storage conditions and choose a product that has a higher commercial value than tilapia, so that results would be more useful to the market. As such, blue crab meat was chosen for the study.; In order to compare the two technologies to a standard quality measurement, blue crab claw meat was sampled over its commercial storage period of 14 days on ice. Total Volatile Base Nitrogen (TVB-N) was used a baseline for measured meat quality and models were generated using data collect from both e-nose and VIS/NIR technologies. E-nose was found to be able to predict the TVB-N level in the meat with an accuracy of less than 5.0 mg TVB-N/100 g. Using visible spectroscopy TVB-N levels were predicted to an accuracy of 4.8 mg TVB-N/100 g. These values were found to be in the same range as that of the ion-specific electrode TVB-N measurements suggesting that these two technologies have the potential to be more accurate with better measurement of the calibration variable.; Since most research has shown a steady decrease in product quality with time, storage time was used a calibration variable. While this does not tie to a specific chemical indicator, time has the advantage of taking into account many different changes that might be missed by any specific indicator. This study investigated both blue crab lump meat and claw meat. E-nose and VIS/NIR readings were taken throughout the storage time of 14 days. These measurements were then used to create a model that could predict total storage time of a meat sample under specific storage conditions. With e-nose technology a standard error of prediction (SEP) was achieved of 2.48 days for claw meat and 2.77 days for lump meat. VIS/NIR spectroscopy yielded significantly better results with 1.31 and 1.11 days for claw and lump meat respectively.; This research shows that both e-nose and VIS/NIR spectroscopy can be used to generate an estimate of fish quality. By being able to model both time and specific indicators, these two technologies show the