Abstract
This paper describes a method of detecting blood spots in brown-shelled eggs using a spectroscopic analysis. The acquired absorbance mode was employed within the visible spectra of eggs to measure spectral data as the explanatory variable. The discriminant analysis based on the Mahalanobis distances was performed. The discriminant function was made by using 120 samples as the calibration set. Then the prediction, whether it was a normal egg or blood-spot one, was practiced by also using 120 samples as the prediction set. As a result, the discrimination ratio for normal eggs was ascertained to be 98.3% whereas that for blood-spotted ones was 93.3%. Moreover, it was clarified that the discrimination ratio for eggs with blood spots in GP center was only 47.8%. The results suggest that the discriminant analysis used in this paper is an effective method for nondestructive detection of brown-shelled eggs that have blood spots.
