Big Cheese Data – Fully automatic on-line 3D NIRS measurement of Cheese production quality

The dairy industry in Denmark is famous for its cheeses. In 2015, the Danish dairies produced 392,000 tons of cheese, of which 85% was exported. Quality control is therefore crucial and the industry would like to upgrade from laboratory-based, at-line single measurements to 100% on-line quality control of all cheeses.

By: Grith Mortensen

On-line quality control of cheese production is, however, extremely challenging due to the cheese size and shape and consequently heterogeneity and quality-gradients within the cheeses. Near infrared spectroscopy has the potential to provide a unique chemical-physical fingerprint of cheese quality and this project will explore the possibilities of using spatially resolved near infrared spectroscopy for on-line quality control.

This project is based on a method in which a near-infrared spectroscopic method is tailored and automated for cheese analysis directly in the cheese production. This allows the cheese producer to generate a very large database of measurements of different cheeses, production fluctuations, seasonal variation, commodity fluctuations and characterization of cheese development during storage. The database can subsequently be used to monitor and characterize new cheeses and describe their expected quality properties after storage. The proposed method is unique as the database is developed organically by continuously including all new measurements and assessments from the on-going cheese production. The system thus includes a self-learning capacity and will continuously improve its predictive capability of the quality of the finished cheese.

The project is unique in that it for the first time will allow 100% production control coverage of cheese production, and map variations in the produced cheeses and production derivations that may occur as a function to time of year, or variations in the supplied raw goods (milk, water).