A Non-Invasive Sensor for Real-Time Evaluation of Milk Quality and Mastitis

Feldman Yuri, HUJI, School of Computer Science and Engineering, Applied Physics



Food Analysis, health, milk classification

Current development stage

TRL4 Technology validated in lab            



  • The yield and the quality of milk in food industry are highly affected by udder health.
  • Bovine mastitis is manifested as an inflammation of the mammary gland and has major cost implications.
  • Most of existing milk monitoring is carried out on samples from the collection tank and not from the individual cow, thus unable to detect whether an individual cow has this disease.
  • Therefore, there is an increasing requirement for real-time milk quality monitoring.
  • Real-time monitoring implies classification of the milk for different end products. Quantities like fat content, lactose and other milk constituents are important to know. In order to answer these needs, an identification of suitable indicators in milk is vital.

Our Innovation

A unique online multi-sensor evaluating in real time, during milking, the milk quality as it is extracted from the cow teat.

  • Continuous monitoring of a milk product by single cow in real time.
  • Early detection of mastitis in the cow udder by monitoring the produced milk.
  • An indication of milk quality in real time by monitoring Caseins, other proteins and fat content of the produced milk.
  • An integral part of a system connecting the milking teat cup to the storage  tank
  • High sensitivity, specificity and reliability.


  • Sub-clinical mastitis is difficult form to detect due to the absence of any visible alterations in the milk or the udder. However, alterations in milk constituents and an elevated somatic cell count (SCC) can be observed
  • There are specific frequency regions that are very sensitive to the presence of high SCC (indicative of mastitis) and Casein content (indicative of the milk quality for drinking or cheese production). Sensors can identify frequency regions in the dielectric spectra.
  • The multi-senor will contain 2 separate sensing units: a low frequency unit (10 Hz to 100 kHz), based on inductance coils and a high frequency unit (up to 2 GHz), based on a reflected microwave.
  • The multi-sensor will detect if the milk is tainted with bacterial infection, its fat content, water content and other milk constituents and will present a profile of the milk's quality. 


Figure 1: Real time milk quality sensor – schematic


Figure 2: The dielectric spectra of pure water compared with a typical raw milk sample at 25oC and 0.1-50GHz frequency range


  • Increase the bovine milk and dairy products yield and quality.
  • Further automation of the dairy production.
  • Contribute to the comprehension of milk’s structure and dynamics.

Contact for more information:

Amichai Baron
VP, Head of Business Development, Agritech & Envir
Contact ME: