Multispectral Portable Fibre-Optic Reflectometer for the Classification of the Origin of Chicken Eggshells in the Case of Mycoplasma synoviae Infections

The proper classification of the origins of food products is a crucial issue all over the world nowadays. In this paper, the authors present a device-a multispectral portable fibre-optic reflectometer and signal processing patch-together with a machine-learning algorithm for the classification of the origins of chicken eggshells in the case of Mycoplasma synoviae infection. The sensor device was developed based on previous studies with a continuous spectrum in transmittance and selected spectral lines in reflectance. In the described case, the sensor is based on the integration of reflected spectral data from short spectral bands from the VIS and NIR region, which are produced by single-colour LEDs and introduced to the sample via a fibre bundle. The measurement is carried out in a sequence, and the reflected signal is pre-processed to be put in the machine learning algorithm. The support vector machine algorithm is used together with three different types of data normalization. The obtained results of the F-score factor for classification of the origins of samples show that the percentages of eggs coming from Mycoplasma synoviae infected hens are up to 87% for white and 96% for brown eggshells.

Automated summary

In this paper, a multispectral portable fibre-optic reflectometer and signal processing patch, along with a machine-learning algorithm, are proposed for classifying the origins of chicken eggshells with Mycoplasma synoviae infection. The results show that the proposed method can accurately classify infected and non-infected eggshells based on spectral data and support vector machine algorithm.