Assessing the Accuracy of Leg Mounted Sensors for Recording Dairy Cow Behavioural Activity at Pasture, in Cubicle Housing and a Straw Yard
1
Animal Health, Behaviour and Welfare Department, Harper Adams University, Newport TF10 8NB, UK
2
Farm Department, Harper Adams University, Newport TF10 8NB, UK
3
Rezatec–Quad One, Becquerel Ave, Didcot OX11 0RA, UK
4
Agriculture and Environment Department, Harper Adams University, Newport TF10 8NB, UK
*
Author to whom correspondence should be addressed.
Animals 2022, 12(5), 638; https://doi.org/10.3390/ani12050638
Submission received: 31 January 2022
/
Revised: 24 February 2022
/
Accepted: 28 February 2022
/
Published: 3 March 2022
(This article belongs to the Special Issue Dairy Cow Behavior, Welfare, and Productivity)
Simple Summary
Sensors attached to the legs of cows are used to record behaviours such as lying times and step count. It is important that the information recorded by sensors is accurate, as changes to behaviour could indicate to the farmer that a cow is sick or in heat. Therefore, this study compared the behaviours recorded by IceQube sensors to those visually observed by humans, when the cows where housed at pasture, in a straw yard and in cubicle housing. The results showed that lying and standing times, the number of times the cows stood up and laid down and the number of steps recorded by the sensors and human observers was similar. Sensors accurately recording cow activity could potentially be worn by cows throughout their lives and could be used to predict and detect sick cows. This could allow the farmer to respond quicker and treat sick cows sooner, increasing animal welfare.
Abstract
The accuracy of CowAlert IceQube sensors (IceRobotics Ltd., Edinburgh, UK) for recording lying duration, standing and lying transitions and number of steps when dairy cows where at pasture, in cubicle housing and in a straw yard, was investigated. Holstein Friesian cows at Harper Adams University, UK were fitted with IceQube sensors; one on the back left (BL) leg and one on the front left (FL) leg. Cows at pasture (n = 48), in cubicle housing (n = 46) and in a straw yard (n = 45) were visually observed. Data were analysed in two stages: (1) an initial exploratory phase determined the correlation level between sensor measurements andvisual observations. Subsequently, (2) a mixed effects modelling framework was used to check whether sensors provide significantly different measures of cow’s activities compared to the observations. Results indicate that lying and standing times are similar between the observed and recorded times, in all three locations. In terms of sensor placement, significant differences were found between the number of steps recorded between BL and FL on straw and pasture, but all other activities were similar, in each location. The accuracy of CowAlert IceQube sensors on the BL leg gives them the potential to be used as lifelong sensors.
Keywords:
dairy cow; accelerometer; validation; pasture; indoor housing; straw yard
