Skip to main content
ARS Home » Midwest Area » East Lansing, Michigan » Sugarbeet and Bean Research » Research » Publications at this Location » Publication #337009
Title: Prediction of canned black bean texture (Phaseolus vulgaris L.) from intact dry seeds using visible/near-infrared spectroscopy and hyperspectral imaging data

Author
item MENDOZA, FERNANDO - Michigan State University
item Cichy, Karen
item SPRAGUE, CHRISTY - Michigan State University
item GOFFNET, AMANDA - Michigan State University
item Lu, Renfu
item KELLY, JAMES - Michigan State University

Submitted to: Journal of the Science of Food and Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/31/2017
Publication Date: 6/5/2017
Citation: Mendoza, F., Cichy, K.A., Sprague, C., Goffnet, A., Lu, R., Kelly, J.D. 2017. Prediction of canned black bean texture (Phaseolus vulgaris L.) from intact dry seeds using visible/near-infrared spectroscopy and hyperspectral imaging data. Journal of the Science of Food and Agriculture. 98(1):283-290.

Interpretive Summary: Texture influences consumer acceptance of canned bean products. Hardness or softness is perceived by consumers visually and when chewing the bean. Currently, an effective pre-canning test to assess canned bean firmness is not available, and instrumental techniques commonly used to measure firmness requires beans to go through the entire canning process prior to texture evaluation. This method is also costly and time consuming, and requires specialized equipment as well as at least 100 g of dry seed per sample. Hence, it would be very useful if bean breeders and processors are able to determine the final firmness of breeding lines more efficiently and earlier in the breeding process when less seed is available. This study explored the feasibility of two nondestructive optical sensing techniques: the conventional visible and near-infrared reflectance spectroscopy (Vis/NIR, spectral information) and hyperspectral imaging (HYPERS, providing both spectral and spatial information simultaneously) for predicting texture of canned black beans from intact dry seeds (i.e., before canning). Bean samples were grown in Michigan (USA) over three field seasons and exhibited a wide phenotypic variability for canned bean texture due to genetic variability and agronomic practice. Overall, better prediction accuracies were obtained for Vis/NIR than for HYPERS. We concluded that Vis/NIRS and HYPERS have great potential for predicting the end-use firmness quality of beans in industrial canning operations, breeding and research facilities; however, based on the results using independent and mixed sets of bean samples, the robustness of the calibration models seems to be affected by the genotypic diversity and distribution of the firmness data used for model building. Hence, calibration models should be periodically maintained and updated with new data.

Technical Abstract: BACKGROUND: Texture is a major quality parameter for the acceptability of canned whole beans. Prior knowledge of this quality trait before processing would be useful to guide variety development by bean breeders and optimize handling protocols by processors. The objective of this study was to evaluate and compare the predictive power of visible and near-infrared reflectance spectroscopy (Vis/NIRS, 400-2,498 nm) and hyperspectral imaging (HYPERS, 400-1,000 nm) techniques for predicting texture of canned black beans from intact dry seeds. Black beans were grown in Michigan (USA) over three field seasons. The samples exhibited phenotypic variability for canned bean texture due to genetic variability and agronomic practice. Spectral preprocessing methods (i.e., smoothing, first and second derivatives, continuous wavelet transform, and two-band ratios), coupled with a feature selection method, were tested for optimizing the prediction accuracy in both techniques based on partial least squares regression (PLSR) models. RESULTS: Vis/NIRS and HYPERS were effective in predicting texture of canned beans using intact dry seeds, as indicated by their correlation coefficients for prediction (R_pred) and standard errors of prediction (SEP). Vis/NIRS was superior (R_pred =0.546–0.923, SEP =7.5–1.9 kg.100g -1) to HYPERS (R_pred =0.401–0.883, SEP =7.6–2.4 kg.100g -1), which is likely due to the wider wavelength range collected in Vis/NIRS. However, a significant improvement was reached in both techniques when the two-band ratios preprocessing method was applied to the data, reducing SEP by at least 10.4% and 16.2% for Vis/NIRS and HYPERS, respectively. Moreover, results from using the combination of the three-season data sets based on the two-band ratios showed that Vis/NIRS (R_pred =0.886, SEP =4.0 kg.100g -1) and HYPERS (R_pred =0.844, SEP =4.6 kg.100g -1) models were consistently successful in predicting texture over a wide range of measurements. CONCLUSION: Vis/NIRS and HYPERS have great potential for predicting the texture of canned beans; the robustness of the calibration models is impacted by genotypic diversity, planting year and phenotypic variability for canned bean texture used for model building, and hence, chemometric models should be maintained and updated with new data.