Standardization of the fourier transform near-infrared reflectance spectroscopy for estimation of some oil quality parameters in mustard (Brassica spp.)

Abstract

The possibility of the application of the fourier transform near-infrared reflectance spectroscopy (NIRS) to the analysis of the selected quality parameters in the mustard oil was followed to determine oil, protein, erucic acid and crude fibre content at the Central Soil Salinity Research Institute, Karnal, India. The samples were analysed by reference methods and by the fourier transformed near infrared (FT-NIR) spectroscope at integrating sphere within reflectance mode in the wavelength range 10 000-4000/cm (1000-2500 nm) with 32 scans. To develop the calibration model for the examined components, the partial least square was used and this model was validated by full cross validation. The coefficients of determination (r2) for intact seeds were 0.907, 0.922, 0.902 and 0.903 for oil, protein, erucic acid and crude fibre content, respectively thus showing that NIRS calibrations are applicable for the estimation of seed quality parameters which is highly desirable in Brassica breeding programs for a quick and non-destructive analysis of oil, protein, erucic acid and crude fibre contents in intact seed.

Keywords

• mustard
• oil content
• near-infrared spectroscopy
• calibration
• validation