Date of Award

2018

Document Type

Master Thesis

Degree Name

Masters of Science (Research)

Department

Institute of Technology Tralee

First Advisor

Dr. Pat Doody

Second Advisor

Dr. Daniel Riordan

Abstract

The weaning of cattle is a process which is known to be labour intensive and to have stressful effects on both cow and calf Common methods used in the weaning process include the temporary removal of a mother from the calf and manual observation and intervention. Early and speedy weaning is known to have a number of benefits, including health benefits for both cow and calf, additional weight gains for the calves as well as reduced labour and feed requirements. The process known as Two-Stage Weaning is recognised to be an effective low-stress approach to weaning in which the calf and cow are allowed to remain together during the weaning process.

This project’s industry partner has developed an innovative device for use as part of the Two-Stage Weaning process which deters calves from suckling and thus speeds up the weaning process. Key to the success of this technology is the ability to predict when suckling is about to commence in order to intervene.

This thesis presents an evaluation of the use of machine learning techniques to detect presuckling indicators in calves. Two potential methods have been developed and are presented, the detection of calf head movement patterns from motion data and the visual detection of the cow’s udder from image data gathered from a neck mounted camera placed on the calf The constraints to choose the most suitable machine learning algorithms for this application are accuracy, low latency and simplicity, which will allow deployment to an embedded processing unit enabling automatic, real-time detection, allowing timely intervention to deter suckling.

Comments

Submitted to Quality and Qualifications Ireland April 2018

Access Level

info:eu-repo/semantics/openAccess

Share

COinS