Skip to main content
ARS Home » Midwest Area » East Lansing, Michigan » Sugarbeet and Bean Research » Research » Publications at this Location » Publication #277612
Title: Prediction of soluble solids content of blueberries using hyperspectral imaging

Author
item LEIVA-VALENZUELA, G - University Of Catolica De Chile
item Lu, Renfu
item AGUILERA, JOSE MIGUEL - University Of Catolica De Chile

Submitted to: Annual Food Engineering Conference
Publication Type: Abstract Only
Publication Acceptance Date: 2/28/2012
Publication Date: 4/2/2012
Citation: Leiva-Valenzuela, G., Lu, R., Aguilera, J. 2012. Prediction of soluble solids content of blueberries using hyperspectral imaging [abstract]. Annual Food Engineering Conference, April 2-4, 2012, Leesburg, Washington. CDROM.

Interpretive Summary:

Technical Abstract: Soluble solids content (SSC) is one of the most important quality parameters for blueberry, and it directly relates to the marketability and shelf life of fresh blueberries. Currently, color imaging technology is being used for color and size grading of fruits like blueberry. The technology is, however, not effective for quality grading of sweetness or SSC. As an emerging technology, hyperspectral imaging generates a spatial map of spectral information, thus providing a useful means for detecting chemical and physical characteristics of food products. A study was carried out to predict the SSC of blueberries using hyperspectral reflectance imaging in the visible and near-infrared region of 500-1,000 nm. Three hundred blueberries were scanned using a laboratory push-broom hyperspectral imaging system to obtain hyperspectral images for each sample at 6 nm increments. Mean spectra were extracted from the hyperspectral images of each blueberry sample. The SSC of blueberries was measured using a digital refractometer. A SSC prediction model was then developed based on 2/3 of the samples using partial least squares method. The model was able to predict the SSC of blueberries with a correlation coefficient of 0.85 and the ratio of sample standard deviation to standard error of validation being equal to 1.85, which suggests the possibility of sorting blueberries into two SSC classes. Research also showed that further improvement in SSC prediction could be achieved through correct identification and removal of the calyx-end area from the blueberry images in the calculation of mean reflectance spectra. Hyperspectral imaging technique can be potentially implemented in packing houses for sorting and grading blueberries to enhance product quality or marketability and shelf life.