Multivariable sensor data to early detect lameness in sheep

Al-Rubaye, Z., Al-Sherbaz, A., McCormick, W. D., Turner, S. J. and Ghendir, S (2016) The use of multivariable sensor data to early detect lameness in sheep. Paper presented to: Sensors in Food and Agriculture, Møller Centre, Churchill College, University of Cambridge, 29-30 November 2016.

Lameness is a clinical symptom referring to locomotion changes that widely differ from normal gait or posture. Lameness has a negative impact on both farm productivity and sheep welfare. The annual loss to the British sheep industry, because of the footrot only ;which is one of the common lameness causes, is estimated by £10 for each ewe. 

Since lameness is often an infectious disease that can be easily spread¬ within the flock, the prior detection of the lame sheep will be expected to decrease the prevalence of lameness and enabling the shepherd to react quickly to the better treatment. 

The prototype sensor has been developed primarily to conduct this research, offering an automatic monitoring of individual sheep to collect behavioural data measurements from a precise sensor that mounted within a neck collar. The sensor variables include 3-axis acceleration, 3-axis Gyroscope, (Roll, Pitch, Heading) angles, longitude, latitude and time. 

The sensor parameters were be used as inputs to data analysis algorithms. The preliminary results from applying pre-existing classification algorithms gave a positive indication for earlier lameness detection, however; the next experiments aim to simplify the process of lameness detection by eliminating the least effective parameters

All views and opinions are the author's and do not necessarily reflected those of any organisation they are associated with. Twitter: @scottturneruon