A camera system that automatically detects lame cows

Lame cows are a major problem for farmers. These animals are in a great deal of pain and consequently produce less milk. Researchers from Wageningen University & Research have placed cameras around Dairy Campus. The system should automatically detect lame cows at the touch of a button. The researchers want to find out what impact locomotion issues have on these animals’ behaviour. Livestock farmers can use the system to take even better care of their animals.

Infectious and non-infectious hoof problems often leave cows lame. Mortellaro's disease, for example, is often caught from wet and dirty floors. Uncomfortable stalls and long waiting times for milking also cause cows to stand for long periods, and this can also cause hoof problems.

Farmers can then intervene quicker and arrange for extra pedicures by hoof trimmers, for example
Claudia Kamphuis

“Lameness is defined as a clinically perceptible deviation from the gait pattern, but subtler changes can also occur. It is important to recognise these abnormalities early on: milk production is affected due to the pain that the animal experiences,” says animal scientist Claudia Kamphuis. “Farmers can then intervene quicker and arrange for extra pedicures by hoof trimmers, for example. Our colleagues are looking at a specific infectious hoof disease, Mortellaro's disease, and we are looking at the gait pattern and its impact on cows’ behaviour.”

The researchers are equipping innovation centre Dairy Campus with cameras and antennae. Cameras in the milking parlour are set up to capture images of their hooves. Cameras set up near the exit of the milking parlour record the gait of cows passing through the walkway. “We are creating an algorithm that detects deviations from the standard gait pattern. To do this, we are using a system that works using artificial intelligence. We indicate a number of points that are important for the gait – the hooves, hip and knees, for example. Then we teach the algorithm what the normal gait for each cow is. If the gait starts to deviate, due to a hoof disorder for example, we can flag it up early on.”

Behavioural data from the living areas

The researchers combine the resulting data with behavioural data from the living areas. “We determine the connections between gait and the cows’ behaviour. There is a possibility that cows that develop locomotion issues at any point behave differently from those that do not. So particular behaviours may be a risk factor for locomotion issues. For example, animals low in the social hierarchy may be at greater risk than those higher up because they have to stand more and are frequently chased away from the main herd.”

Animals low in the social hierarchy may be at greater risk than those higher up
Claudia Kamphuis

To combine the video and behavioural data, the researchers have fitted the cows with collars that contain sensors. They have also installed cameras in one of the living areas. Antennae in the enclosures pick up signals from the sensors, allowing them to monitor what all the cows are up to: where the cows are walking, how active they are and what behaviour they are displaying, such as standing up or lying down. The cameras are used to both validate sensor information and generate information. In short: how is the animal behaving? “We can continuously monitor animals this way. By using AI, we can obtain information that we simply did not have access to five years ago. This means we can detect problems early on.”

Control room at Dairy Campus

According to Kamphuis, the trick is to analyse and store all the collected data intelligently. The researchers want to do this by creating a modern data warehouse. “We are thinking about modern data warehouse infrastructure for linking and tracking data intelligently. The process involves large amounts of information and is multidimensional and complex. Researchers must also be able to access the material they need for their research easily. So, how long we should keep what data is also important.”

At Wageningen University & Research, we will continue to focus on innovative and new analysis techniques
Claudia Kamphuis

Ultimately, with systems like these, farmers can identify cows with locomotion issues earlier on and provide appropriate care. Breeders can also use the data for breeding by selecting for reduced susceptibility to hoof disorders. “The Dairy Campus and the video tracking system are a catalyst for further research. We are going to develop specific knowledge on the prevention of locomotion problems, for example, through housing and feeding measures, because we follow the animals consistently over a period time with this system. Colleagues can use these techniques too, to look at individual feed intake, for example. At Wageningen University & Research, we will continue to focus on innovative and new analysis techniques.”