Influence of the homogenizing grade and meathematical treatment on the determination of ground beef components with near infrared reflectance spectroscopy

식품의 근적외선 반사분광분석법에서 균질의 정도가 흡광도에 미치는 영향 및 수학적 처리방법에 관한 연구

This study was conducted to determine the effect of the homogenizing grade of sample on absorbance of near infrared reflectance spectrophotometer with which chemical compositions of food were rapidly and effectively analyzed. By the mathematical treatment of absorbance values standard error of prediction was reduced as follows. 1. The absorbance values of various samples ground for the same periods of time were calibrated before or after treatment with first or second derivative in an attempt to accurately predict the components of samples ground for the different periods of time. The standard error of prediction for moisture content were 1.478%, 0.658% and 0.580%, respectively, those for fat content 0.949%, 0.637% and 0.527%, respectively, and those for protein content 0.514%, 0.493% and 0.394%, respectively. Calibration of absorbance values after second derivative treatment showed the highest accuracy in predicting sample components. 2. The absorbance values of various samples ground for the different periods of time were calibrated before or after treatment with first or second derivative in order to accurately predict the components of samples ground for the different periods of time. The standard error of prediction for moisture content were 1.026%, 0.589% and 0.568%, respectively, and those for protein content 0.860%, 0.557% and 0.399%, respectively. The standard error of prediction were lower in the order of calibrations before and after first and second derivative treatments. As a result, calibration of absorbance values after second derivative treatment showed higher accuracy regardless of grinding time of samples.

신속간편법으로서 식품의 성분함량 측정에 이용되는 근적외 반사분광분석에서 문제가 되고 있는 시료의 균질도의 차이가 흡광도에 미치는 영향을 조사하고 흡광도로 나타난 수치를 수학적으로 처리함으로써 예측의 오차는 다음과 같이 감소하였다. 1. 세절도가 일정한 시료의 흡광도를 무변형, 1차 미분 및 2차 미분하여 calibration한 후 세절도가 다양한 시료의 성분을 예측한 결과 예측오차(standard error of prediction)는 수분함량에 있어서 1.478%, 0.658% 및 0.580%, 지방함량의 예측오차 0.949%, 0.637%, 및 0.527%이었으며 단백질 측정시 각각 0.514%, 0.493% 및 0.394%로 2차 미분의 예측오차가 모든 성분측정에서 가장 낮음을 알 수 있었다. 2. 세절도가 다양한 시료의 무변형, 1차 미분 및 2차 미분처리한 calibration을 이용하여 세절도가 다양한 시료의 성분을 예측한 결과 수분함량 예측오차는 각각 1.026%, 0.589% 및 0.568%, 지방함량 예측오차는 각각 0.828%, 0.639% 및 0.602% 이었으며 단백질함량 예측오차는 0.860%, 0.557% 및 0.399%로서 역시 2차 미분, 1차 미분, 무변형의 순으로 정확성이 높았다. 결과를 종합할 때 흡광도를 수학적으로 처리하여 calibration에 이용하므로서 calibration시료의 세절도에 관계없이 측정결과의 정확성은 높아졌다.