An Identification of raw milk samples obtained from estrus and non-estrus dairy cows using Micro-NIR Spectrometer

Abstract

This project is a feasibility study of an identification of raw milk obtained from estrous and non-estrous dairy cows using near-infrared spectroscopy combing with principal component analysis techniques. The samples were measured by a micro-NIR spectrometer(950-1650nm). The use of traditional tools that need to be installed on cow and chemical methods can be challenging. However, using near-infrared spectroscopy had more advantage such as no effect on cows and fast. The raw milk samples used in the study came from four Holstein dairy cows that were sampled in the morning and evening for 40 days. The samples were scanned for collecting NIR spectra and then the principal component analysis (PCA) was used to analyze milk spectra across the MATLAB program. The results showed that the spectral pretreatment by the Moving Average Smoothing + Standard Normal Variate method was clearly classified into two groups: 1) Proestrus and 2) Estrus and Metestrus. The study shows that it was possible to use NIR spectroscopy to classify milk samples from estrus and non-estrus cows. However, future studies should investigate additional factors may affecting the classification model, such as study more algorithms and create model using important individual. Wavelength, in order to get a more accurate.