Light Optimization for an LED-Based Candling System and Detection Combined with Egg Parameters for Discrimination of Fertility

Highlights An LED-based candling system was designed to automatically judge egg fertility. LEDs of six different colors were tested for light optimization. Relative absorptivity (RA) was negatively correlated with the L* value and positively correlated with the a* value. The results combined with egg parameters produced higher accuracy than the results without egg parameters. Abstract. Removal of infertile eggs could increase the efficiency of commercial hatcheries by saving space, ensuring the hatching rate, and decreasing costs. In this study, an LED-based candling system was designed to automatically judge egg fertility by measuring photo-resistance, which was converted into the relative absorptivity (RA) of light through the egg. In the first experiment, 85 eggs were used to test LEDs of six different colors (red, yellow, green, blue, warm white, and cold white) during days 0 to 13 of incubation. In the second experiment, 170 eggs were detected with warm white LEDs during days 0 to 10 of incubation to further study whether egg parameters (weight, dimensions, and shell color) affected the accuracy of logistic regression. The detection results of the first experiment indicated that the RA of fertile eggs increased rapidly throughout the first 13 days of incubation, while the RA of infertile eggs increased little. Comparing the detection results for all LEDs, it was found that the warm white LEDs achieved the best results, with an accuracy of 94.9% on day 7. The second experiment showed that the RA was negatively correlated with the L* value (lightness or darkness) and positively correlated with the a* value (color in the red or green direction). Furthermore, on most days of incubation, the results of logistic regression combined with the egg parameters produced higher accuracy than the results without egg parameters. The accuracy was 100% on day 10. This study provides an automatic and non-destructive method to discriminate fertile eggs from infertile eggs for current hatcheries.