Considering pre-processing of accelerometer signal recorded with sensor fixed on dairy cows is a way to improve the classification of behaviours
| Autor: | Riaboff, Lucile, Bedere, Nicolas, Couvreur, Sébastien, Aubin, Sébastien, goumand, E., Magnier, J., Madouasse, Aurélien, Chauvin, Alain, Plantier, Guy |
|---|---|
| Přispěvatelé: | Terrena Innovation, ESEO-GSII (GSII), ESEO-Tech, Université Bretagne Loire (UBL)-Université Bretagne Loire (UBL), Laboratoire d'Acoustique de l'Université du Mans (LAUM), Le Mans Université (UM)-Centre National de la Recherche Scientifique (CNRS), URSE, Ecole Supérieure d’Agricultures, Univ. Bretagne Loire, 49000, Angers, France, Biologie, Epidémiologie et analyse de risque en Santé Animale (BIOEPAR), Institut National de la Recherche Agronomique (INRA), Centre National de la Recherche Scientifique (CNRS)-Le Mans Université (UM), ProdInra, Archive Ouverte |
| Jazyk: | angličtina |
| Rok vydání: | 2019 |
| Předmět: |
[SDV]Life Sciences [q-bio]
Cow cascade de classifieurs Pasture characteristics [SDV.SA.ZOO]Life Sciences [q-bio]/Agricultural sciences/Zootechny accelerométrie comportement animal [SDV] Life Sciences [q-bio] analyse du signal [STAT.ML]Statistics [stat]/Machine Learning [stat.ML] classification vache laitière Decision tree accelerometry Behaviour dairy cows Global positioning system [SPI.SIGNAL]Engineering Sciences [physics]/Signal and Image processing |
| Zdroj: | 9. European Conference on Precision Livestock Farming (ECPLF) 9. European Conference on Precision Livestock Farming (ECPLF), Aug 2019, Cork, Ireland 27. General meeting of the European Grassland Federation (EGF) 27. General meeting of the European Grassland Federation (EGF), Jun 2018, Cork, Ireland. pp.807-809 |
| Popis: | The relationship between animal behaviour and pasture characteristics needs to be explored to facilitatepasture management. Using Global Positioning System (GPS) data, this paper aims to study the influenceof pasture characteristics and time of day on (1) cow distribution in the paddock and (2) the mainbehaviour inferred in each area. Fourteen Holstein cows were tracked with GPS receivers mounted oncollars for an entire day and were observed simultaneously. A vegetation cover characterisation (floralspecies and dry matter) and sward height measurements were carried out before and after grazing. Eachof these measurements was also GPS located. A heterogeneous distribution of cows inside the paddockwas observed. Using distances and turning angles, main behaviours (grazing, resting and walking) werepredicted using a decision tree. Cows spent significantly more time resting in the afternoon than in themorning and frequently rested near the drinking trough. A greater difference in sward height betweenentry and exit of cows was observed in areas where resting was the main behaviour. These findingsshowed that cow time-budget can be studied efficiently without laborious observation using GPS-locatedresources and GPS data. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|