Online quality analysis of raw milk: potential of miniature spectrometers

Over the past 50 years, genetic selection and improved feed and management practices resulted in an increased milk production per cow lactation. As these modern cows are prone to production-related disorders, they need to be monitored closely to guarantee animal health and welfare. On the other hand, the increasing size of dairy farms results in less time available for each individual animal. Therefore, tools are needed to provide reliable and useful information on individual cows and help the dairy farmer to optimize the animal management while reducing the workload. Milk contains valuable information on the metabolic and nutritional status of dairy cows. Therefore, regular analysis of the produced milk is an efficient way to monitor cow health and welfare. Frequent milk analysis is only feasible if it is performed on the farm with a minimal investment of labor and resources. Different studies, both in the lab and on the farm, indicate that near infrared (NIR) spectroscopy holds the potential for rapid, non-destructive and on-line analysis of the raw milk composition. Nevertheless, commercial NIR detectors are typically costly (> 10k€) and relatively large, not allowing for implementation in existing milking systems. In this study, we evaluated the ability of miniature NIR sensors to analyze raw milk and predict the fat, protein and lactose concentration. These NIR sensors use alternative approaches for wavelength scanning, which enables compact sensor packaging and fast signal collection. The performances of the miniature spectrometers will be compared against benchtop (diode array) NIR spectrometers in an on-farm validation study.