Veterinärmedizinische Universität Wien Forschungsinformationssystem VetDoc

Grafischer Link zur Startseite der Vetmeduni Vienna

Gewählte Publikation:

Publikationstyp: Zeitschriftenaufsatz
Dokumentart: Originalarbeit

Publikationsjahr: 2018

AutorInnen: Roland, L; Lidauer, L; Sattlecker, G; Kickinger, F; Auer, W; Sturm, V; Efrosinin, D; Drillich, M; Iwersen, M

Titel: Monitoring drinking behavior in bucket-fed dairy calves using an ear-attached tri-axial accelerometer: A pilot study.

Quelle: Computers and Electronics in Agriculture 2018; 145: 298-301



Autor/innen der Vetmeduni Vienna:

Drillich Marc,
Iwersen Michael,

Beteiligte Vetmed-Organisationseinheiten
Bestandsbetreuung bei Wiederkäuern,


Abstract:
Acceleration sensors allow a reduction of time-consuming visual observation. The drinking behavior of bucket fed calves has not been monitored automatically yet, although sufficient milk intake is essential for calves' health and growth. The objectives of this pilot study were (1) to evaluate the technical and mathematical feasibility of using an acceleration sensor to detect drinking events in bucket-fed dairy calves, (2) to develop an algorithm for an acceleration sensor (SMARTBOW ear tag, Smartbow GmbH, Weibern, Austria) for monitoring drinking behavior in bucket-fed dairy calves, and (3) to validate the SMARTBOW sensor for monitoring drinking events in bucket-fed dairy calves to observations from video recordings. Three preweaned dairy calves were equipped with ear-tag accelerometers. Calves were housed in individual pens and fed milk from a teat-bucket twice a day. Acceleration data were collected and calf behavior was video-recorded for 5 d for 24 h d(-1). Based on a training data set, an algorithm was developed to predict drinking events. Further 15 d of data were generated by simulation. Video recordings were used to analyze whether drinking events (n = 174) were predicted correctly for the complete data set. Sensitivity (82.9%), specificity (96.9%), and accuracy (96.2%) were good, but precision (60.4%) was not yet optimal. Cohen's Kappa (0.68) indicated substantial agreement between sensor and video analysis. More research based on a larger number of animals with the aim to optimize the underlying algorithm and to further increase sensitivity and precision is planned.


© Veterinärmedizinische Universität Wien Hilfe und Downloads