Non-intrusive automated measurement of dairy cow body condition using 3D video

Hansen, Mark F; Smith, M; Smith, L; Hales, I

Abstract

Regular scoring of a dairy herd in terms of various physical metrics such as Body Condition Score (BCS), mobility and weight are essential for maintaining high animal welfare. This paper presents preliminary results of an automated system capable of nonintrusively measuring BCS automatically as the cow walks uninhibited beneath a 3D camera. The system uses a ’rolling ball’ algorithm on the depth map which simulates how well a ball of a set radius fits the surface. In this way a measure of angularity is generated which is shown to be inversely related to BCS on 95 cows. The measurements
are shown to be highly repeatable with 14 out of 15 cows being scored within one quarter BCS score repeatedly and seven of those being scored within an eighth of a BCS score.

Citation

Hansen, M. F., Smith, M., Smith, L., & Hales, I. (2015, September). Non-intrusive automated measurement of dairy cow body condition using 3D video. Presented at British Machine Vision Conference - Workshop of Machine Vision and Animal Behaviour

• Presentation Conference Type: Speech
• Conference Name: British Machine Vision Conference - Workshop of Machine Vision and Animal Behaviour
• Start Date: Sep 10, 2015
• End Date: Sep 10, 2015
• Publication Date: Jan 1, 2015
• Peer Reviewed: Peer Reviewed
• Keywords: dairy cows, 3D video, measurement, non-intrusive
• Additional Information: Title of Conference or Conference Proceedings : British Machine Vision Conference - Workshop of Machine Vision and Animal Behaviour; Corporate Creators : Duncan Forbes