|DONG, JUN - China Agricultural University
|DONG, XIAOGUANG - China Agricultural University
|LI, YANLEI - China Agricultural University
|PENG, YANKUN - China Agricultural University
|Chao, Kuanglin - Kevin Chao
|GAO, CUIYING - Inner Mongolia Saifeiya Agricultural Science And Technology Development Co, Ltd
|TANG, XIUYING - China Agricultural University
Submitted to: Computers and Electronics in Agriculture
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/28/2019
Publication Date: 6/4/2019
Citation: Dong, J., Dong, X., Li, Y., Peng, Y., Chao, K., Gao, C., Tang, X. 2019. Identification of unfertilized duck eggs before hatching using visible/near infrared transmittance spectroscopy. Computers and Electronics in Agriculture. 157:471-478.
Interpretive Summary: Traditional candling method is widely used to evaluate fertility of hatching duck eggs. However, this method is time consuming, labor intensive and subject to human error. Undetected unfertilized eggs in incubation can burst and easily contaminate other eggs. It increases food safety risk to consumers. Rapid and non-destructive detection methods are needed that can be used on-site at field production to detect unfertilized eggs before hatching. This study developed a spectroscopic measurement technique to acquire spectral data from duck eggs. Statistical models were developed to predict unfertilized eggs. The best model was used in 667 duck eggs to detect unfertilized eggs with an accuracy of 94.7%. The results show that the spectroscopy method combined with mathematical models can be used as a rapid and non-destructive method to detect unfertilized duck eggs. The method developed in this study is useful to inspect unfertilized eggs that could greatly benefit the poultry industry and consumers.
Technical Abstract: Identification of unfertilized duck eggs before hatching is difficult to accomplish in the poultry industry by the traditional candling method. Visible/near infrared (VIS/NIR) transmittance spectroscopy was used as a rapid and non-destructive technique to distinguish the unfertilized duck eggs before hatching. Transmittance spectra of duck eggs laid by the same flock of ducks were acquired on site at a poultry farm. Statistical models were validated by the internal leave-one-out cross-validation (LOOCV) and external validation. The prediction model developed using the Naive Bayes (NB) method achieved the best performance. The classification accuracy for calibration set and validation set was 94.5% and 94.7%, respectively. The results show that the VIS/NIR transmittance spectroscopy is an effective method for rapid and non-destructive detection of unfertilized duck eggs.