Assessment of internal quality of blueberries using hyperspectral transmittance and reflectance images with whole spectra or selected wavelengths

Authors: LEIVA-VALENZUELA, GABRIEL - Catholic University - Chile; Lu, Renfu; AGUILERA, JOSE MIGUEL - Catholic University - Chile

Submitted to: Innovative Food Science and Emerging Technologies
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/5/2014
Publication Date: 2/16/2014
Citation: Leiva-Valenzuela, G., Lu, R., Aguilera, J. 2014. Assessment of internal quality of blueberries using hyperspectral transmittance and reflectance images with whole spectra or selected wavelengths. Innovative Food Science and Emerging Technologies. DOI: 10.1016/j.ifset.2014.02.006.

Interpretive Summary: Good quality blueberries should be firm and sweet. Soft blueberries are a major quality concern because they are susceptible to mold development and enzymatic browning and have a shorter shelf life. It is thus desirable or even necessary to sort and grade individual blueberries for soluble solids content (SSC), indicative of sweetness, and firmness to assure consistent product quality and consumer satisfaction. In this research, hyperspectral imaging technique was used for predicting the SSC and firmness of blueberries. The technique was implemented in both reflectance and transmittance modes for acquiring three-dimensional spectral images of blueberries for the visible and near-infrared spectral region of 500-1,000 nm. Four hundred twenty blueberries were imaged for three fruit orientations (i.e., stem end, calyx end and equator) to evaluate the effect of fruit orientation and sensing mode on the prediction of SSC and firmness using the hyperspectral imaging technique. The SSC and firmness of each berry were measured using standard, destructive refractometric and compression methods. Mathematical techniques were applied to extract spectral features from the acquired hyperspectral images for prediction of blueberry SSC and firmness. Results showed that reflectance mode was better than transmittance mode in predicting both SSC and firmness, with the best correlation coefficient of 0.90 for SSC and 0.78 for firmness. Fruit orientation had small or negligible effect on SSC and firmness measurement. Using a few select wavelengths resulted in about 5% lower prediction accuracies in SSC and firmness, compared with using whole spectra. Hyperspectral imaging technique could be implemented with a few select wavelengths for sorting and grading blueberries into two quality grades of SSC and/or firmness.

Technical Abstract: Hyperspectral imaging has been used in previous studies for assessing firmness and soluble solids content of fresh fruit. To apply this technique for automatic sorting and grading of blueberries, it is necessary to investigate different sensing modes (i.e., reflectance and transmittance), evaluate the effect of fruit orientation on fruit quality prediction, and develop robust prediction models with fewer wavelengths. In this study, a hyperspectral imaging system was used to acquire reflectance and transmittance images from 420 blueberries in three fruit orientations (i.e., stem end, calyx end and equator) for the spectral region of 400-1,000 nm. Mean spectra were extracted from the regions of interest for the hyperspectral images of each blueberry. Calibration models for firmness index (FI) and soluble solids content (SSC) were developed using partial least squares regression for the reflectance and transmittance spectra as well as their combined data. Further, interval partial least squares (iPLS) regression with 10 different intervals of nine wavelengths was used to reduce the spectral dimensionality. Overall, reflectance gave better results (the best correlation for prediction or R of 0.90 for SSC and 0.78 for FI) than transmittance (R of 0.76 for SSC and 0.64 for FI). For reflectance, FI and SSC predictions for the stem-end orientation were generally better than for the other two orientations, while fruit orientation had little or insignificant effect on transmittance predictions. Combination of reflectance and transmittance spectra did not yield improved prediction results for both FI and SSC. The prediction errors for iPLS, on average, increased by about 5%, compared to PLS for the whole spectra. The research demonstrated that it is feasible to implement hyperspectral imaging technique for sorting blueberries for SSC and possibly firmness, using appropriate wavelengths.