Artifical Intelligence to deliver step change in pig husbandry

A new project is underway to develop an automated service for farmers to provide 365 day surveillance of their pigs. The FarmSense project has been funded by DEFRA in conjunction with Innovate UK to improve and monitor pig welfare and provides an automated surveillance service with a unique disease detection system, creating daily reports and alerts on performance and behaviour, enabling farmers to monitor their livestock remotely.

Led by Innovent Technology and costing £785,554 , FarmSense is an intelligent user-friendly platform which uses state-of-the-art image and sensor technologies placed on farms, combined with artificial intelligence to support farmers in optimising livestock production whilst assuring the highest animal welfare standards.

FarmSense’s smart monitoring system continuously analyses animal growth, behaviour and gas profiles along with the animal’s day and night patterns. The AI system learns how to automatically detect any changes in pattern indicating problems such as:

  • early disease onset
  • tail biting
  • abnormal eating or drinking behaviours.

This intelligent user-friendly platform brings together state-of-the-art imaging and Volatile Organic Compound (VOC) sensing technologies with artificial intelligence (AI)and machine learning to support pig producers, stockpersons and their advisors.

The system uses a method, developed by Innovent Technology Ltd, of smart 3D cameras that monitor animal growth and behaviours, such as tail-biting, to provide the farmer with insights into the behaviour and weight of animals, avoiding unnecessary and stressful interventions to the pigs. Innovent’s David Barclay said, “It’s critical to invest in technology for the pork industry, projects such as FarmSense allow consortiums like ours to push the boundaries and bring innovation into the sector. Our aim is always to bring value to the producer, this system will allow us to identify where potential is being lost in finishing, and monitor the outcome of change to address that loss.”

Volatile Organic Compounds are emitted from living organisms and are different according to the health status of the animal – such as when an infection or disease is present. Cambridgeshire based Roboscientific has developed a new generation of VOC sensors which ‘sniff’ the air surrounding the pigs for specific VOC digital fingerprints associated with a disease to identify in real time when there is an outbreak in the herd. Commenting on the new project, Roboscientific CEO, Stan Curtis observed “We are very excited to be a part of this project. We have already proven our ability to spot diseases in chickens at an early stage, allowing poultry farmers to react quickly and efficiently to minimise the impact a disease could have on a flock. We believe that this project will allow us to develop a similar Early Disease Detection method for pigs and help farmers to improve the way that they manage their herds.”

Zoetis, the leading global animal heath company, will provide the integrated platform through which the farmer will be notified in real time of anything unusual or alarming occurring within a herd to allow for early corrective action and fast, effective farm management decisions to be made. “Disease detection without action still means costly impacts to herd health and productivity. Given the current, and exacerbating labour challenges that farms will face, providing that timely information will be critical for UK farmers to remain competitive” Gonzalo Rivero, Pig and Poultry Director at Zoetis UK, explained.

Trials to develop the new system are taking place at Scotland’s Rural College (SRUC) research farm near Edinburgh and overseen by Rick D’Eath, a Senior Research scientist in Animal Behaviour with over 23 years research experience in the pig sector, who said “We have worked with Innovent before to develop and validate tail posture for tail biting detection, and the combined package with the VOC element, offers the chance of an integrated solution. We are pleased to provide a test-bed for developing and testing this technology at our research farm.”

Follow-on commercial scale trials will take place within both outdoor and indoor production systems owned by Cranswick plc. Commenting on the trials, Ash Gilman, Director of Agricultural Strategy at Cranswick plc said: “We installed a small number of Innovent’s 3D cameras in one of our finishing sites in August 2020, as I was keen to learn exactly what they were capable of. Adding the next level of surveillance and analytics to our businesses is vital. Early disease detection, scientific weight selection without intervention, and welfare indicators need even more focus to ensure we are both sustainable and profitable in future. I am extremely keen to see this technology developed and scaled up commercially as part of the project.”

Data analysis to develop integrated Machine Learning and AI methodology is being undertaken by the Data Innovation Hub for Animal Health (DIHAH) at the University of Surrey. Team leader and Associate Professor Kevin Wells commented, “FarmSense perfectly fits the intended use of our DIHAH platform: to make use of and add value to animal health data. It draws on our experience on animal health analytics and Surrey’s internationally renowned AI capabilities.”

The project is scheduled to run until April 2024 when the ambition is to deliver a commercial version of the Farmsense Monitor improving animal welfare, increasing efficiency and improving quality in the pork supply chain.

Images: Roboscientific’s non-invasive automated Early Disease Detection Farm Monitor (prototype) and Overhead 3D pig behaviour and performance camera