We use cookies on our website. Some are necessary for the operation of the website. You can also allow cookies for statistical purposes. You can adjust the data protection settings or agree to all cookies directly.
Data Protection Declaration
Imprint
Collaborative project: A monitoring tool for animal welfare in dairy cows. Subproject 3 (CowAlarm)
Project
Project code: 2815710215
Contract period: 16.02.2017
- 15.04.2020
Budget: 91,480 Euro
Purpose of research: Applied research
Keywords: digitalisation
The aim of the "CowAlarm" project was to develop a monitoring system for identifying conspicuous behavioural deviations (sickness behavior) in dairy cows. For this purpose, a practical farm was equipped with commercially available sensor systems that provided serial information on activity, resting behaviour, feed intake and ruminating behaviour. This information was analysed and grouped using principal component analysis and partial least squares regression. Multivariate cusum charts and sequential decision trees (XGBoost) were used for conspicuous/nonconspicuous classification. Good accuracies were obtained with both approaches (sensitivity: 70%, specificity; 94%, 6% false positive alarms). Cusum Charts can be easily implemented in the management software, XGBoost algorithms require a central solution due to the higher computing capacity.
Section overview
Subjects
- Agricultural Technology Animal Production
- Animal health
- Computer science
- Farm animal behaviour
