Building a Better Cow: Project focuses on improving efficiency, profitability, wellness

MADISON, Wis. — The goal of the Building a Better Cow project is to improve dairy cow productivity, profitability and wellness.

“This project is focused on building a better cow using 3D imaging for phenotyping of conformation traits,” said Jeffrey Bewley, executive director of genetic programs and innovation for the Holstein Association USA.

“Artificial intelligence is changing a lot of things in our world and a lot of things in the dairy industry,” he said during a presentation at the Knowledge Nook held at the World Dairy Expo. “I am very excited about what is coming.”

Conformation, Bewley said, has an important impact on the lifetime performance of an animal.

“We did a study a few years ago where we compared classification scores and individual linear traits to lifetime performance,” he said. “What we saw was lifetime production increased in each classification score category.”

For example, Bewley said, cows that got a score of 85 to 89 in their first lactation produced about 30,000 pounds more milk than cows that were lacking in conformation.

“So, conformation does matter for selection of animals within the dairy breed,” the executive director said.

“We think everything is in place — the tools, technology and the need to transition how we look at evaluating cow conformation,” he said. “We have spent the last couple of years developing an innovative solution designed to advance cow conformation evaluation into a revolutionary era.”

The Holstein association had applied for a patent for the Building a Better Cow system.

“The goals of our project are to conduct conformation trait evaluation reliably and consistently and also record some management traits like body condition score and some locomotion indicators,” Bewley said. “This stationary system is primarily targeted at large dairy farms to supplement, but not replace classifications.”

The system uses three time-of-flight cameras that are strategically placed in the exit alley from the milking parlor. The cameras provide high resolution 3D images of each cow as it passes through the system.

“At the current time, we are measuring 26 traits, including body condition score, a few locomotion indicators and all of our linear traits,” Bewley said.

Time-of-flight cameras measure the depth across an object by calculating the time it takes for the light to get to the object and back to the camera.

“Unlike conventional cameras, these provide us with more accurate information about the object we’re looking at,” Bewley said.

“We capture 3D images of the cow as she leaves the parlor, we use artificial intelligence through techniques of machine learning to process the information from that image, and then we calculate the traits using basic principles of geometry,” he explained.

“After we calculate those traits, we deliver the traits in a spreadsheet format to integrate the data into herd management software,” he said.

One camera is positioned above the cow, one is behind the cow to get udder traits and the third camera points at the side of the cow.

The first prototype of the Building a Better Cow system was installed at Western Kentucky University.

“We have two other systems out there — one is on a 2,200-cow dairy milking three times a day and the third system is at a 600-cow dairy also milking three times per day,” Bewley said.

The Holstein board has been actively engaged with this project from the start.

“I think that’s one of the things that makes this unique — our board has helped design the parameters,” Bewley said. “And provided us many good insights to make this something we think will be useful.”

For body traits, the system measures standard traits such as stature, body depth and chest width.

“We have multiple things we’re looking at for feet and legs,” Bewley said. “With the rear camera we can look at udder width, front and rear teat placement.”

“We’ve taken a little different approach for locomotion because we want to be able to understand the difference in movement and how we could bring this into new traits to describe how animals move,” he said. “We are not looking at mobility in terms of if she is lame or not, but how the animal moves.”

For example, there are multiple gait measurements.

“Instead of just a locomotion score, we are looking at swinging in or swinging out behavior,” Bewley said. “And tracking behavior to see if the rear foot lands in the same place as the front foot when the cow is walking.”

The Holstein association is also considering additional traits.

“Right now, I have a list of about 110 traits that we could measure on a cow,” the executive director said. “We think there’s a lot of potential for adding new traits, but for a while it’s going to be 26 traits.”

Although the primary focus for the system has been for Holstein cows, Bewley said, there is potential for it to be extended to other dairy breeds as well as for beef breeds.

“Another area I have a lot excitement for is thinking about how we can use this data to do more predictive analytics by combining the data we get from the imaging system with genomic data and milk data,” he said.

“So, we can do a better job of predicting things like longevity, lameness risk and lifetime profitability to change how we make decisions about individual cows.”

As the Building a Better Cow system continues through the research and development phase, Bewley said, the plan is to install it on additional farms in 2026.

“We hope to be ready to sell this system in 2027 or 2028,” he said.