Improving Automatic Detection of Abnormal Milk

When cows are milked with automatic milking systems or in high capacity milking parlors, milk quality cannot be checked sufficiently without sensors. The most commonly used parameters are electrical conductivity (EC), color of the milk and milk yield, mostly determined at quarter level. In order to get information from these measurements, relatively straightforward algorithms are used before alerts lists are generated. Subsequently, the farmer has to check or pay extra attention to the alerted cows or quarters. A major complaint, however, is that these alerts list produce far too many false-positive alerts.