Studies on nondestructive detection of abnormall eggs using optical sensing

Chicken eggs are now widely used almost exclusively for food, and more than 100 million eggs are distributed each day in Japan. At Grading and Packing (GP) center in Japan, inspectors identify abnormal eggs by candling with naked eye. Though some GP centers are using the egg candlers, the detection accuracy is not so high. In this study, nondestructive detection methods for abnormal eggs are discussed by using the optical sensing technology. 1. The detection of blood spots in white-shelled eggs A method by which blood spots can be detected is examined by using visible spectroscopy. The determination is made using the wavelength absorbance data for the hemoglobin in the blood spots of the eggs. As a result, the discrimination ratio for normal eggs and eggs containing blood spots is 100% and 83.0%, respectively. PLS regression analysis is performed with the explanatory variable within the spectral range from 500 to 600 nm. When predictions are made for unexamined specimens from three different production areas, the discrimination ratio for normal eggs is 100%, and that for eggs with blood spots is 96.8%. The results suggest that PLS regression is an effective method for predicting eggs with blood spots. 2. The detection of eggs with blood spots for brown-shelled eggs A method of detecting blood spots in brown-shelled eggs is examined by using a spectroscopic analysis. The acquired absorbance mode is employed within the visible spectra of eggs to measure spectral data as the explanatory variable. The discriminant analysis based on the Mahalanobis distances is performed. The discriminant function is made by using 120 samples as the calibration set. Then the prediction, whether it is a normal egg or blood-spot one, is practiced by using 120 samples as the prediction set. As a result, the discrimination ratio for normal eggs ascertained to be 98.3% whereas that for blood-spotted ones 93.3%. Moreover, it is clarified that the discrimination ratio for eggs with blood spots in GP center is only 47.8%. The results suggest that the discriminant analysis used in this study is an effective method for nondestructive detection of brown-shelled eggs that have blood spots. 3. The detection of various abnormal eggs A method of detecting various abnormal eggs; (outlying yellow, lacerated yellow, no yellow, eggs with air spot) is examined by using spectroscopic analysis. As a result, a clear difference is observed between the normal eggs and eggs with outlying yellow at the wavelength of 560nm. As for eggs with lacerated yellow, a clear difference is observed between the two at the wavelength of 615nm. As for no yellow eggs, it is clarified that the absorbance of no yellow eggs is low in the spectral range of 550 to 800nm. As for eggs with air spots, a clear difference is observed between the two in the spectral range over 750nm. From the result above, it is concluded that the optical sensing technology is an effective method to detect various abnormal eggs nondestructively