Sensor-based approaches to monitoring the behaviour, health, and welfare of dairy calves
Currently, the detection of ill health in UK farmed calves is based on intermittent visual observation which is subjective and poorly accurate. Sensor-based monitoring may offer an improved alternative. For example, sensors could be used to monitor behaviour and detect signs of ill health in calves....
| Main Author: | |
|---|---|
| Format: | Thesis (University of Nottingham only) |
| Language: | English |
| Published: |
2023
|
| Subjects: | |
| Online Access: | https://eprints.nottingham.ac.uk/73153/ |
| _version_ | 1848800767918473216 |
|---|---|
| author | Carslake, Charles |
| author_facet | Carslake, Charles |
| author_sort | Carslake, Charles |
| building | Nottingham Research Data Repository |
| collection | Online Access |
| description | Currently, the detection of ill health in UK farmed calves is based on intermittent visual observation which is subjective and poorly accurate. Sensor-based monitoring may offer an improved alternative. For example, sensors could be used to monitor behaviour and detect signs of ill health in calves. However, substantial individual variation exists for many behaviours, the extent of which is poorly understood in calves. Here, within- and between- individual variation in calf feeding behaviours are quantified using data from computerised milk feeders. Results show that substantial, temporally stable individual differences exist. In addition, the average behavioural expression of two distinct feeding behaviours were positively and significantly correlated and the between-individual differences observed were shown to be consistent over time and context, and to be associated with weight gain. This improves our understanding of normal variation in calf feeding behaviour, which could be helpful in detecting potential behavioural changes indicative of ill health. Machine learning models were trained and tested using feeding data from computerised milk feeders to detect ill health. In a separate study, a similar methodology was used to detect ill health using reticulo-rumen temperature boluses. Results indicate low and moderate predictive performance, respectively. Study limitations and areas for future research are discussed. Finally, the development of novel technologies to enable a more holistic approach to behavioural monitoring in calves is explored. Results show that signals from a single collar-based sensor can be used to accurately detect nine different behaviours as well as to quantify rarely occurring behaviours, such as locomotor play. Quantifying play behaviour could provide a useful indicator of positive welfare in calves. It is also shown that these behaviours can be detected using computer vision, but that further work is needed to enable generalisation to new camera angles and scenes. Overall, this thesis highlights the potential of sensor-based technologies to improve our understanding of behavioural variation in calves, as well as to monitor a greatly more diverse range of behaviours than previously attempted. It is hoped that this work will contribute towards the improvement of health and welfare in calves. |
| first_indexed | 2025-11-14T20:56:48Z |
| format | Thesis (University of Nottingham only) |
| id | nottingham-73153 |
| institution | University of Nottingham Malaysia Campus |
| institution_category | Local University |
| language | English |
| last_indexed | 2025-11-14T20:56:48Z |
| publishDate | 2023 |
| recordtype | eprints |
| repository_type | Digital Repository |
| spelling | nottingham-731532023-07-31T04:40:53Z https://eprints.nottingham.ac.uk/73153/ Sensor-based approaches to monitoring the behaviour, health, and welfare of dairy calves Carslake, Charles Currently, the detection of ill health in UK farmed calves is based on intermittent visual observation which is subjective and poorly accurate. Sensor-based monitoring may offer an improved alternative. For example, sensors could be used to monitor behaviour and detect signs of ill health in calves. However, substantial individual variation exists for many behaviours, the extent of which is poorly understood in calves. Here, within- and between- individual variation in calf feeding behaviours are quantified using data from computerised milk feeders. Results show that substantial, temporally stable individual differences exist. In addition, the average behavioural expression of two distinct feeding behaviours were positively and significantly correlated and the between-individual differences observed were shown to be consistent over time and context, and to be associated with weight gain. This improves our understanding of normal variation in calf feeding behaviour, which could be helpful in detecting potential behavioural changes indicative of ill health. Machine learning models were trained and tested using feeding data from computerised milk feeders to detect ill health. In a separate study, a similar methodology was used to detect ill health using reticulo-rumen temperature boluses. Results indicate low and moderate predictive performance, respectively. Study limitations and areas for future research are discussed. Finally, the development of novel technologies to enable a more holistic approach to behavioural monitoring in calves is explored. Results show that signals from a single collar-based sensor can be used to accurately detect nine different behaviours as well as to quantify rarely occurring behaviours, such as locomotor play. Quantifying play behaviour could provide a useful indicator of positive welfare in calves. It is also shown that these behaviours can be detected using computer vision, but that further work is needed to enable generalisation to new camera angles and scenes. Overall, this thesis highlights the potential of sensor-based technologies to improve our understanding of behavioural variation in calves, as well as to monitor a greatly more diverse range of behaviours than previously attempted. It is hoped that this work will contribute towards the improvement of health and welfare in calves. 2023-07-31 Thesis (University of Nottingham only) NonPeerReviewed application/pdf en cc_by https://eprints.nottingham.ac.uk/73153/1/Charles%20Carslake%20%E2%80%93%2014343522%20%E2%80%93%20thesis.pdf Carslake, Charles (2023) Sensor-based approaches to monitoring the behaviour, health, and welfare of dairy calves. PhD thesis, University of Nottingham. Precision livestock farming; Animal welfare; Animal personality; Animal behaviour |
| spellingShingle | Precision livestock farming; Animal welfare; Animal personality; Animal behaviour Carslake, Charles Sensor-based approaches to monitoring the behaviour, health, and welfare of dairy calves |
| title | Sensor-based approaches to monitoring the behaviour, health, and welfare of dairy calves |
| title_full | Sensor-based approaches to monitoring the behaviour, health, and welfare of dairy calves |
| title_fullStr | Sensor-based approaches to monitoring the behaviour, health, and welfare of dairy calves |
| title_full_unstemmed | Sensor-based approaches to monitoring the behaviour, health, and welfare of dairy calves |
| title_short | Sensor-based approaches to monitoring the behaviour, health, and welfare of dairy calves |
| title_sort | sensor-based approaches to monitoring the behaviour, health, and welfare of dairy calves |
| topic | Precision livestock farming; Animal welfare; Animal personality; Animal behaviour |
| url | https://eprints.nottingham.ac.uk/73153/ |