Hyperspectral imaging for non-destructive prediction of total nitrogen concentration in almond kernels

T. Gama, H.M. Wallace, S.J. Trueman, I. Tahmasbian, S.H. Bai
There is increasing awareness of the need to consume high-quality foods because of health concerns. Food safety and health awareness campaigns have provided an impetus for non-destructive and real-time methods for food quality assessment. Total nitrogen is used as an indicator of crude protein content in foods and we examined the potential of hyperspectral imaging to predict total nitrogen concentration in four brands of almonds purchased from commercial retailers. A hyperspectral imaging system in the wavelength range 400-1000 nm was used in the study. A partial linear squares regression (PLSR) model was developed, which predicted total nitrogen concentration with a determination coefficient (R2p) of 0.82 and a root mean error square of calibration (RMSEC) of 0.16. These results indicated that hyperspectral imaging has great potential to predict total nitrogen concentration of almond kernels.
Gama, T., Wallace, H.M., Trueman, S.J., Tahmasbian, I. and Bai, S.H. (2018). Hyperspectral imaging for non-destructive prediction of total nitrogen concentration in almond kernels. Acta Hortic. 1219, 259-264
DOI: 10.17660/ActaHortic.2018.1219.40
https://doi.org/10.17660/ActaHortic.2018.1219.40
almond, crude protein, food quality, nutritional composition, nuts, rapid assessment
English

Acta Horticulturae