Machine vision technology for automatic behaviour assessment of young bulls in pens

Archive ouverte

Cheype, A. | Manceau, J. | Gauthier, V. | Dugué, C. | Merle, L.A. | Delahaye, Quentin | Boivin, Xavier | Mindus, Claire

Edité par CCSD -

International audience. Changes in animal's behaviour are good indicators of health and welfare variations. However, human observation is time-consuming and labour-intensive. Development of video technology and image processing may offer the opportunity for a better prevention by detecting behavioural issues continuously, automatically, and therefore at an early stage. We are developing deep learning algorithms to analyse routinely the behaviour of young bulls. This study evaluates performances of algorithms developed to automatically detect the different activities of bulls on images. Bulls originating from 2 different breeds were housed accordingly to the standard management conditions of their respective stations (Pôle de Lanaud, Ferme des Etablières). Two cameras 2 D color were installed above each pen of bulls with different angular views. Postures and 5 behaviours were labelled on 1108 images extracted from the videos. Then 2 sets were used to evaluate the performances of the algorithm which is an object detection model that uses convolutional neural networks to detect and classify objects in an image: 790 images for training and 175 for validation. The classification precision and sensitivity of bulls were promising with respectively 97% and 78% for standing, 99% and 46% for lying, 62% and 72% for standing up, 64% and 70% for lying down, 94% and 93% for eating, 63% and 77% for drinking, 59% and 72% for moving. This project BeBoP contribute to the current need for on-farm, operational behavioural indicators that can be easily used to assess not only the individual welfare but also the welfare of the whole group.

• Description
• Sujet/Public
• Outil
• Outil d'anticipation
• Mémoire et thèse
• Gestion