Light optimization for A LED-based candling system and detection of egg fertility

Abstract. Detection of egg fertility will increase efficiency in commercial hatcheries such as saving space, ensuring hatching rate and decreasing costs. Therefore, a LED-based candling system was designed to detect fertilized eggs. Totally, 6 kinds of LED (Red, Yellow, Green, Blue White-2700K and White-5900K) boards were tested in the first experiment for light optimization. During the first 13 days of incubation, eggs sampled from an egg farm were detected every day, then broken out and categorized as fertilized, unfertilized and dead embryos. Discriminant analysis was applied to discriminate unfertilized eggs by using the relative absorption rate (RAR) data. The continuous measurement data indicated that the average RAR of fertilized eggs increased rapidly while the average RAR of unfertilized eggs increased a little. Compared the detection results of 6 LED boards, the results of using two LED (Red and White-2700K) boards was the best with the accuracy of 94.9% at day 7. Another detection experiment, where the weight, size and pigment intensity were measured, were employed. The result showed that combined with the physical indexes of eggs, the accuracy reached 95.9% at day 7, 96.5% at day 8, 99.4% at day 9 relatively. This study provided an automatically and non-destructive method to distinguish fertilized and unfertilized for current hatcheries.