Near infrared spectroscopy: a rapid and economical method for predicting protein in milled rice
2007
Garcia, G.DG. | Mamucod, H.F. | Camus, R.A. | Manaois, R.V., Philippine Rice Research Inst., Science City of Munoz, 3119, Nueva Ecija (Philippines)
Among the essential parameters being measured to assess rice quality are the physicochemical properties which include moisture, amylose and crude protein (CP) content. These are routinely assessed through wet-laboratory methods which unfortunately are laborious, time-consuming, costly, and pose harm to the environment. The use of Near Infrared (NIR) spectroscopy holds promise as a faster, more accurate, and economical alternative method to the conventional physicochemical analyses. Spectra of milled whole grains of 108 rice cultivars harvested from 2005 wet and 2006 dry seasons were collected using the near-infrared regions (850-1048 nm) of the 1241 Grain Analyzer. The reference CP values of the samples, ranging from 3.86-10.29%, were obtained using kjeldahl method. Population structuring technique called Principal Component Analysis was done to reduce the spectra and select the samples best for developing the mathematical model. The best model developed for CP prediction named R1120706 used SNV and Detrend scatter correction with 1, 4, 4, 1 mathematical treatment. The statistical data of the best CP model shows that a good prediction model was developed having a small standard error of calibration (SEC) and standard error of cross-validation (SECV) of 0.25 and 0.26, respectively. In addition, the standard error of prediction (SEP) value of 0.29 is within the standard of a perfect equation (SEP = 1.2 to 1.5 x SEL) with 1.2X of its standard error in the laboratory (SEL=0.55). Employing the developed model R112 0706 will allow for the use of NIR to rapidly and nondestructively measure the CP content in milled rice without the costly and time-consuming constraints of normal laboratory operations.
显示更多 [+] 显示较少 [-]