AI to improve animal monitoring in slaughterhouses


AI to improve animal monitoring in slaughterhouses

Connecting stakeholders for impact

In recent years, large slaughterhouses in the Netherlands have installed camera surveillance systems to better monitor how animals are handled. In practice only a random selection of the many hours of video footages is viewed.

That is why Deloitte developed an innovative, AI-based camera model in collaboration with the animal welfare organisations De Dierenbescherming and Eyes on Animals, and meat producer Vion. With AI, movements of people, animals and objects and how they interact are monitored. This enables slaughterhouse employees to better detect and address animal handling issues.

“Two years ago I read an article by two Deloitte colleagues about innovation in the food industry. They wrote about the possible application of video analytics in slaughterhouses. And that got me thinking”, says Carlos Morales, Partner at Deloitte Consulting. "Why not use innovative technology to make a positive impact where you might not expect it, but is much-needed?" Morales investigated whether the idea was feasible and together with colleague Jasmijn Korver proposed it to the Deloitte Impact Foundation (DIF). The DIF dedicates time and expertise on a pro bono basis to support initiatives which impact society. The Foundation’s board was enthusiastic about the idea, and funded hours for the pilot phase of this project.

Connect for impact

“There were no reusable examples available, so our AI4Animals team had to start from scratch”, Morales explains. “We came into contact with meat producer Vion through animal welfare organisation De Dierenbescherming. A collaboration between animal welfare organisations and slaughterhouses is not as unlikely as you might think. There is a common denominator: the need to better monitor animals. By jointly focusing on this, a unique collaboration was created. De Dierenbescherming and Eyes on Animals were able to tell us all about animal welfare, what causes stress and what can go wrong. Vion made video material available and gave us a tour around the slaughterhouse to see with our own eyes how to improve camera surveillance. The team then reviewed hours video recording to develop and feed the AI system.”

Recognise animals, people and objects

Up until now, slaughterhouses would view a random selection of the daily video recordings. Sjors Broersen, Manager at Deloitte Technology Consulting: “This meant that most of the footages remained unseen and potential animal welfare problems stayed invisible. Using AI and machine learning, we taught the system to recognise animals, people and objects, such as electric stunning devices and paddles, and to analyse movement and interaction. If a pig is staying behind, it might have been injured at the farm or during transportation. If there are too many animals in a cluster, this may indicate a bottleneck in the process that can cause unnecessary stress for the animals. The system also detects potential signs of life and consciousness after animals have been stunned and bleed. Video fragments of these kinds of situations are collected in a dashboard for review so that corrective actions can be taken.”

Innovative technology

It sounds like a simple and logical system, but a lot of time and effort has gone into the development. Broersen: “The team analysed many hours of footage and we used state of the art technology to develop this new video software. The system also detects potential signs of life or consciousness after animals have been stunned and bled. Moreover, the system continues to learn based on the feedback from the slaughterhouse officers.” 

“This is just the beginning”, Morales continues. “We are now also applying AI to monitor cows and we have started a pilot for chicken. A lot more is possible on the basis of video and audio analytics. For the future, we are thinking of better monitoring other parts in the process, such transport. And possibly the recognition of other sources of suffering such as fevers and wounds. Traditional cameras should not be the industry standard, intelligent cameras should be the industry standard. That is our goal. For the Netherlands and, of course, beyond. If organisations are interested to team up with us to develop and apply innovative technology to better monitor animals, they are very much welcome”, Morales concludes.

Deloitte Future of Food

In Deloitte’s Future of Food platform we aim to accelerate the transformation of the global food system to one that’s more responsible, sustainable and healthy. We do this by connecting the ecosystem, sharing inspiration, and by helping companies in their transformation to a future proof food business.

For more information on smart animal welfare, please contact Carlos Morales or Sjors Broersen from the AI4Animals team. Reach out to Randy Jagt to discuss the Future of Food from a comprehensive perspective. If you want to stay informed, and get invited to our webinars and round tables, sign up for our Future of Food community.

Are you ready for the Future of Food?

Read our insights, reports and impact stories

Get inspired on the Future of Food
Did you find this useful?