In order to provide an alternative to day-old chick culling in the layer hatcheries, a noninvasive method for egg sexing is required at an early stage of incubation before onset of embryo sensitivity. Fluorescence and Raman spectroscopy of blood offers the potential for precise and contactless in ovo sex determination of the domestic chicken (Gallus gallus f. dom.) eggs already during the fourth incubation day. However, such kind of optical spectroscopy requires a window in the egg shell, is thus invasive to the embryo and leads to decreased hatching rates. Here, we show that near infrared Raman and fluorescence spectroscopy can be performed on perfused extraembryonic vessels while leaving the inner egg shell membrane intact. Sparing the shell membrane makes the measurement minimally invasive, so that the sexing procedure does not affect hatching rates. We analyze the effect of the membrane above the vessels on fluorescence signal intensity and on Raman spectrum of blood, and propose a correction method to compensate for it. After compensation, we attain a correct sexing rate above 90% by applying supervised classification of spectra. Therefore, this approach offers the best premises towards practical deployment in the hatcheries.

In order to provide an alternative to day-old chick culling in the layer hatcheries, a noninvasive method for egg sexing is required at an early stage of incubation before onset of embryo sensitivity. Fluorescence and Raman spectroscopy of blood offers the potential for precise and contactless in ovo sex determination of the domestic chicken (Gallus gallus f. dom.) eggs already during the fourth incubation day. However, such kind of optical spectroscopy requires a window in the egg shell, is thus invasive to the embryo and leads to decreased hatching rates. Here, we show that near infrared Raman and fluorescence spectroscopy can be performed on perfused extraembryonic vessels while leaving the inner egg shell membrane intact. Sparing the shell membrane makes the measurement minimally invasive, so that the sexing procedure does not affect hatching rates. We analyze the effect of the membrane above the vessels on fluorescence signal intensity and on Raman spectrum of blood, and propose a correction method to compensate for it. After compensation, we attain a correct sexing rate above 90% by applying supervised classification of spectra. Therefore, this approach offers the best premises towards practical deployment in the hatcheries.