Preliminary study on the use of near infrared hyperspectral imaging for quantitation and localisation of total glucosinolates in freeze-dried broccoli

Abstract

1The use of hyperspectral imaging to (a) quantify and (b) localise total glucosinolates in 2 florets of a single broccoli species has been examined. Two different spectral regions 3 (vis-NIR and NIR), a number of spectral pre-treatments and different mask development 4 strategies were studied to develop the quantitative models. These models were then 5 applied to freeze-dried slices of broccoli to identify regions within individual florets 6 which were rich in glucosinolates. The procedure demonstrates potential for the 7 quantitative screening and localisation of total glucosinolates in broccoli using the 950-8 1650 nm wavelength range. These compounds were mainly located in the external part 9 of florets. 10

Detection of postharvest quality loss in broccoli by means of non-colorimetric reflection spectroscopy and hyperspectral imaging

“…A similar steep slope was observed by Hernández-Hierro et al (2014) in broccoli spectra. However, the remaining spectral pattern obtained by Hernández-Hierro et al (2014) in the visible range (450-680 nm) was rather flat. In the present work, a peak was observed at 550 nm (Fig.…”

“…From 680 to 750 nm, there was a steep rise in reflectance, as it is characteristic for spectra of green plant parts (Knipling, 1970). A similar steep slope was observed by Hernández-Hierro et al (2014) in broccoli spectra. However, the remaining spectral pattern obtained by Hernández-Hierro et al (2014) in the visible range (450-680 nm) was rather flat.…”

“…The 2nd derivative of reflectance spectra from 380 to 1000 nm and the change in chlorophyll concentration were the input factors in an artificial neuronal network (Hosaka et al, 2012). In a preliminary study, Hernández-Hierro et al (2014) related NIR wavelength segments to glucosinolate content in broccoli powders and then estimated its spatial distribution in broccoli slices by hyperspectral imaging. Senescenceinduced changes of reflectance spectra have been reported in different fruits and vegetables without anthocyanin-synthesis, e.g.…”

“…Concerning broccoli, this may be a knowledge gap: Due to the complex morphology, senescence changes within individual broccoli curds can be expected to occur inhomogenously on the curd surface based on single floret buds (Kieffer et al, 1998;Tian et al, 1995). Hernández-Hierro et al (2014) applied image analysis for determination of spatial glucosinolate distribution. However, they used a longitudinal slice of broccoli instead of the curd surface, where senescence changes normally occur.…”

Recent Advances in De‐Noising Methods and Their Applications in Hyperspectral Image Processing for the Food Industry

“…As a result, PLS regression models with SNV pretreated spectra gave the highest of 0.8308 and the lowest root mean square error of prediction set (RMSEP) of 0.243 for inspecting the total viable counts (TVC) content in pork meat. More recently, SNV has been employed as a preprocessing approach to remove slope variation and to correct for light scatter in many endeavors, including the quantification of total glucosinolates in freeze‐dried broccoli (Hernández‐Hierro and others ) and the determination of microbiological loads in beef fillets (Panagou and others ). Huang and others () exploited the potential of using a near‐infrared (NIR) HSI system for identifying milk quality.…”