Click on “Download PDF” for the PDF version or on the title for the HTML version.

If you are not an ASABE member or if your employer has not arranged for access to the full-text, Click here for options.

Time Series Analysis of Inline Sensor Milk Data in Mastitis Detection

Published by the American Society of Agricultural and Biological Engineers, St. Joseph, Michigan

Citation:  2010 Pittsburgh, Pennsylvania, June 20 - June 23, 2010  1008953.(doi:10.13031/2013.29771)
Authors:   Xiangyu Song, Rik van der Tol
Keywords:   Clinical mastitis, Time series, Electrical conductivity, Colors

Clinical mastitis (CM) in dairy cows is a major animal welfare issue. It also concerns milk quality, food safety and farm profit in dairy industry. The aim of this research was to combine quarter-based milk data recorded by a new developed in-line sensor Milk Quality Control TM (MQC, Lely Industries N. V., Maassluis, the Netherlands) in a times series analysis to detect CM automatically with a high sensitivity (SN) and minimum specificity (SP) of 99%. So far data includes 658 cows (Holstein- Friesian, Montbliarde and its crossbreeds) with 135,638 times of milking coming from 10 Lely Astronaut A3 TM milking robots (Lely Industries N. V., Maassluis, the Netherlands) on 5 commercial farms. The reference was 36 detected CM cases reported by farmers. Time series data electrical conductivity (EC) and light transmittance of milk (red, green, blue and infrared were used as the inputs of the detection model. The number of milking in history used for calculating CM alerts and the time window of reported CM cases were the model variables. After testing the CM detection model with different variable combinations, best results were obtained with 3 previous milking and a 5-day time window. The SN and SP reached 75.00% and 99.55%, respectively. It was demonstrated the feasibility of the inline milk related data used by the time series model to detect CM in an AMS.

(Download PDF)    (Export to EndNotes)