Know more

abstract :

In recent years, many monitoring tools have been developed to assist farmers in detecting physiological (calvings, heats) or pathological (mastitis) events. The aim of this Ph.D. thesis was to generate knowledge on the interest of general activity and rumination time measured by a sensor designed for dairy cow monitoring (HR-Tag, SCR Engineers Ltd, Israel), or of alarms generated based on these data.
First of all, the interest of rumination time for dry matter intake prediction during lactation was investigated. The addition of rumination time to an existing dry matter intake prediction model (NRC 2001) significantly improved the prediction, but the gain was not sufficient to enable accurate intake phenotyping. Then, the informative value of alarms generated from activity and rumination data for the detection of health disorders was evaluated. Three complementary reference methods were used successively to determine the health status of the cows: detection of health disorders by visual appraisal by farm staff, identification of major drops in milk production and clinical veterinarian examination of the animals.
The results of this work show that even if heat alarms triggered by the monitoring device are reliable, the performance of the algorithms developed are not good enough for farmers to delegate entirely health disorders detection to the alarm system.

Key words :

Dairy cattle, health monitoring, precision dairy farming, epidemiology, algorithms, cusum, rumination, activity