Automatic individual pig detection and tracking in pig farms

Over one billion pigs are farmed globally each year and can experience a range of health and welfare problems such as respiratory and enteric diseases and tail biting outbreaks. It is a time consuming and labour-intensive task for farmers and vets to monitor their animals day-to-day. Additionally, animals often change their behaviours in the presence of humans, meaning that we don’t always get a clear picture of the true activity levels in a pen. Being able to remotely monitor the pigs behaviour means that we can quantify normal movement patterns and therefore know when a pig is behaving in an unusual way. Deviations from baseline movement, such as less time spent eating and drinking, can signify potential sickness and can aid farmers in knowing where they should focus their efforts.

The first step in achieving a health and welfare warning system is to develop the technology which allows for individual pigs to be detected and tracked in a group. The Pig Sustain project, led by Prof Lisa Collins at the University of Leeds, is doing just that. In recently published work authored by Dr Lei Zhang (University of Lincoln), the team show that their setup of CCTV cameras and a custom algorithm using advanced computer vision and machine learning technologies can detect and track pigs with over 97% precision and 95% accuracy. Crucially, the system works without the need to physically mark or tag the pigs, meaning that farmers do not need to change their normal practices to use the technology.

The team are currently expanding their system to work on larger groups of pigs and future plans are to make the algorithm transferable across different pig farming systems (such as on a straw floors).

PigSustain is funded through the Global Food Security’s ‘Resilience of the UK Food System Programme’, with support from BBSRC, ESRC, NERC and Scottish Government.

Zhang, L.; Gray, H.; Ye, X.; Collins, L.; Allinson, N. Automatic Individual Pig Detection and Tracking in Pig Farms. Sensors 2019, 19, 1188.

Full article / DOI can be found here.