AI System Helps Keep Cows Sound on Their Feet and Producing Milk
Farmers in Russia's Kaliningrad region have introduced the Russian digital platform TsPS Obyektiv (Objective) at the Kashtanovka livestock complex, cutting the prevalence of lameness in the herd by half.
Catching a disease early is far better than treating it after it has progressed. This medical principle applies to animals just as well as it does to people. Early and accurate diagnosis can extend an animal's productive life and increase farm revenue. On large farms, digital predictive video analytics systems are particularly effective at doing this. One such system has been deployed at the Kashtanovka livestock complex in Russia's Kaliningrad region.
Neural Network to Assess 3,900 Cows
The farm belongs to the Zalesye agricultural holding and is considered one of the largest in Russia, with a herd of 3,900 cattle. Monitoring the health of every animal in a herd of this size is a major challenge. Lameness is particularly difficult to detect. Sometimes called the “silent killer” of dairy cows, it often goes unnoticed in its early stages and develops gradually, but can lead to serious, sometimes irreversible, losses.
Visually assessing every cow in a large herd is difficult and sometimes impossible. This was the main reason for introducing the TsPS Obyektiv digital platform, which helps livestock specialists and veterinarians assess the animals' health. Computer vision systems scan the cows as they pass through a monitoring area after milking. A neural network analyzes their gait by tracking 38 key anatomical points. It assesses stride symmetry and length, joint flexion angles, and changes in spinal curvature.
“When the cows leave the milking area, they pass through special gates that read their transponders. Each animal has its own unique ID. At that point, cameras record how the cows move through the gates. If a cow is limping, the algorithm flags it for examination,” said Vadim Mikhailov, head of IT at the Zalesye agricultural holding.
The system's assessment accuracy is 98%. Based on the data, AI automatically assigns each animal an individual locomotion score. If it detects suspicious signs, the platform sends a preliminary diagnosis to the veterinarian. The cow can then be placed under closer observation even before clinical signs of disease appear, allowing treatment to begin earlier.
Improving the Health of Russia's Dairy Herds
The technology has demonstrated its effectiveness in a short time. Installation of the TsPS Obyektiv platform took just 14 days, and commissioning did not disrupt the farm's daily operations. The work included 3D mapping of the sensor cameras to the farm layout, including its robotic milking carousel.
Over the following 90 days, the system helped cut the number of lame animals in half. This reduced spending on treatment. And because lameness can reduce milk yields by 10–20%, the farm's productivity also increased.
The computer vision and AI system has thus created an additional tool for veterinary oversight. Its key advantage is that it can monitor a large herd automatically, around the clock. Accurate assessments of animal health make it possible not only to start treating detected conditions promptly but also to take preventive measures.
Russian developers have gained a new market segment in which demand is expected for computer vision and video analytics systems; motion-recognition algorithms; integration of AI with veterinary information systems; databases for collecting, storing and analyzing livestock health data; and farm management software.
Industrial Livestock Farming Powered by IT
Russia is now developing its own expertise in digital cattle health monitoring and precision livestock farming. For example, students at Saint Petersburg Electrotechnical University LETI are developing a Smart Cow Monitoring system to assess cow health. The system records everything happening inside the barn or out on pasture in real time, after which AI evaluates the animals' condition.
Researchers at the Project Institute of Digital Transformation of the Russian State Agrarian University – Moscow Timiryazev Agricultural Academy have developed a virtual simulator for training users to perform linear conformation assessments of cattle. Another project from the university is an intelligent system for evaluating and analyzing linear conformation traits, with the aim of detecting lameness at an early stage.
As a result, IT is becoming a tool for day-to-day management on modern livestock farms. Automation now extends beyond feeding, milking and herd accounting to include monitoring animal health. By combining computer vision systems, thermal imaging cameras, sensors and neural networks, farms can build comprehensive monitoring systems that track the health of individual animals and help predict potential diseases. This will make it possible to forecast farm productivity with a high degree of accuracy, encouraging investment in the sector and, in turn, increasing supplies of milk and meat to the food market.
