Skip to main content

Artificial Intelligence for Livestock Health Management.

Article Category

Article available in the folowing languages:

Cameras may soon be watching the herd to detect problems earlier

Serket was the Egyptian goddess of nature, animals, medicine, and magic. Now, a company by the same name has developed a cloud-based system exemplifying all those traits to detect potential health issues in livestock before they spread.

Digital Economy
Food and Natural Resources

Agriculture is poised for a revolution in the use of information and communications technology and the EU-funded SERKET project is leading the charge. The project is transforming the business of agriculture with an easy to use artificial intelligence (AI)-based platform for livestock monitoring. Addressing a critical need SERKET scientists spoke with pig farmers in Hungary, Italy, and the Netherlands, where they found an unsustainably high mortality rate of livestock. Approximately 3 % to 5 % die between weaning and slaughter (the so-called fattening period), a number that jumps to 16 % if the time from birth to slaughter is taken into account. In addition to illness, aggression and open wounds are a common source of infections. Given the important shortage of labour in agriculture, there is not enough time to carefully inspect livestock on a regular basis to identify problems before they get out of control. According to project coordinator and SERKET CEO Kristóf Nagy, “Group treatment - antibiotics often added to water or food for the entire herd - happens because farmers and workers discover infections late in the game when much of the herd is already infected. This increases antibiotic resistance as well as the levels of antibiotics in water used for agriculture and drinking. Our studies showed that 80 % of problems can be detected visually. That is currently not possible due to the labour shortage on farms.” Off-the-shelf security cameras for 24/7 surveillance SERKET has developed a sophisticated yet simple-to-use system enabling round-the-clock vigilance and AI-based data analysis to detect abnormal behaviours and alert farmers early. The only hardware needed are inexpensive, commercially available security cameras. The software is designed to detect changes in five different measures of behaviour: activity level, aggression, eating, drinking, and social behaviour. An alert tells the farmer to check on the herd in a certain location and suggests potential remedies. Feedback from the farmer on actual conditions helps the AI deep-learning algorithms improve their performance. The agricultural industry wants and needs Big Data As Nagy explains, “Until now, the fattening period of livestock production was a black box. The animals were weighed at birth and at slaughter. What happened in between was a mystery. Mortality data on the country level were numbers typically without reasons.” SERKET scientists are currently working with about 40 farms ranging from those having about 1 500 animals to large integrators having millions of pigs. In addition, SERKET is running 5 pilot studies paid for by big feed companies and systems integrators. Nagy reflects, “These people found us through our website and offered to pay for the pilot studies at their facilities. The agricultural industry wants and needs a solution.” While data is the key to the revolution, Nagy emphasises “We offer the farmers cost-effectiveness. With our platform, they can reduce mortality, reduce the number of workers needed, optimise feeding, and reduce antibiotic usage. The end result is a reduction in costs.” In addition to farmers, SERKET’s collaborative, interactive data-collection platform will benefit animal breeders, university researchers and government policy makers.


SERKET, farmer, antibiotic, livestock, data, agriculture, herd, animals, mortality, pilot studies, camera, pig, labour shortage, artificial intelligence (AI), agricultural industry, infection, aggression, surveillance, smartphone, policy, breeders,

Discover other articles in the same domain of application