The Creation of the FT-NIR Calibration for the Determination of the Amount of Corn Grain in Concentrated Feed

Abstract

The research is financed by BAPK of Afyon Kocatepe University. Project number: 16 SAGBIL.17. The autors thank to AKU-BAPK.

Abstract This study was carried out to determine the proportion of corn grain used in laying hen concantrate feeds quantitatively by means of near infrared spectroscopy (NIR) device. For this purpose, different laying hen concentrate feed consisting of 60 pieces of 1 kg mixture was formed by increasing the concentration of 1% in each ration from 1% to 60%. Functional NIR spectra of feed mixtures were obtained at wavelengths of 10000-4000 nm / cm. The spectra were evaluated by using PLS (Partial Least Square) method on the second derivative. In the normalization study, SNV (Standard Normal Variate) method was applied. In addition, data obtained from first order (1st Derivation B Cap 5 Points Gap 2) were prepared for regression. Linear regression was applied to the normalized spectra by subtracting the Outlier values from the calibration set and calibration quality parameters were revealed. At this stage, the standard deviations of the R2 value, validation and calibration set were calculated. According to the reflections, the graphs obtained with Regression Coefficients were taken. In addition, the sum of the squares of the estimation residual error (V-Set PRESS) of the validation set was also revealed. Outlier values were determined by the program and were not included in the study due to lowering the calibration quality. In the research, R value was determined as 0.9985, R2 value, 0.9970 and standard error value as 0.9352. The estimated residual error of the validation set obtained was around 14 with the help of V-Set PRESS (Principal Components). The obtained regression consistency was obtained by dividing the standard error value (SEC) of the calibration to the standard error of estimation (SEP). The consistency value was determined between 80 and 110. These results suggest that NIR or FT-NIR spectroscopy can be used to predict the extent to which the corn grain is quantitatively introduced in laying hen feed mixtures. As a result, it was concluded that the percentage of each feed raw material entering the concentrate feed mixes can be quantitatively determined using FT-NIR spectroscopy as a fast and practical method.

Keywords: Corn grain, laying hen feeds, FT-NIR, quantitative

DOI: 10.7176/JSTR/5-8-08