34373

Hyperspectral Imaging to predict shelf-life of ready-to-eat Quinoa sprouts

Favorite this paper

Chenopodium quinoa is an annual plant mainly cultivated for his seeds which have a high content of gluten-free proteins, mineral salts, B vitamins and anti-oxidant substances. Quinoa is commercialised and consumed in different forms: seeds, flour and recently sprouts. The germination of the seeds boosts, in fact, their nutritional value in antioxidants which are also more easily absorbed by human intestine. Therefore the development of a ready-to-eat product based on fresh quinoa sprouts, which include softened seeds, young leaves and stems, may be highly desirable. However, because of its heterogeneity, quinoa sprouts are not well suited for a traditional quality evaluation based on colorimetric or even spectroscopic approaches: sample preparation and representativeness of the acquisition would in fact be critical aspects to solve. Techniques such as Visible or Near Infrared Spectroscopy allow a rapid quality assessment of the product, as it is already well established for other foodstuff. Hyperspectral Image Analysis is the best response to the need of keeping both a fast, not invasive analytical method and a limited – if any – sample preparation. In this work, the potentiality of this approach, together with Multivariate Analysis Methods to extract and understand the relevant information, was tested on Quinoa sprouts in order to evaluate the effect of storage time and different oxygen level of the storage atmosphere. Quinoa samples were stored in controlled atmosphere under 3 levels of oxygen (5%, 10% and 21% as control in air) and evaluated after 0, 2, 5 and 7 days of storage; three replicates were considered for each oxygen condition. Hyperspectral Images were acquired for each combination of storage time and oxygen level, with an hyperspectral scanner (version 1.4, DV srl, Padova, Italy) equipped with 2 spectrographs one in the range of 400 to 1000 and the second in the range 900-1700 nm (spectral resolution of 5 nm and spatial resolution 315dpi). Images were processed by means of Principal Components Analysis (PCA) in order to implement a fast method based on scores thresholding for background removal and extract only the pixels corresponding to the sample. The average spectrum over the selected pixels was used to evaluate, both by means of PCA and ANOVA-Simultaneous Component Analysis (ASCA), if the storage treatment, time and their interaction could have a significant effect on signal modifications. Results showed that storage time had a statistically significant effect on NIR signal, and in particular, that the changes in the signal were mostly related to a decrease of absorption in the region centred at 1405 nm, which can be related to O-H first overtone absorption. These changes can be referred to a modification in the water content of the samples with storage time, which is in good agreement with physical measurements assessed on the samples, such as weight loss. Modifications appeared more evident when comparing the shortest storage time (2 days) with the longest one (7 days). These results are indicating that Hyperspectral Imaging may be a fast and not invasive technique for monitoring quality of Quinoa sprouts.