Tracking cattle can be complicated, but a trained eye and the right algorithm can keep producers in touch with their barns
A startup based out of Belfast, Northern Ireland, developed a low cost and hands-free technology that enables dairy operators to monitor their cattle.
CattleEye uses data from just about any affordable camera to run through its cattle identification and health-monitoring system.
“We build a number of neural networks, artificial intelligence networks, to take 2D security camera footage from cows as they enter and exit the milking parlour. We run these through neural networks, and then we derive insights about the cow,” said Terry Canning, chief executive officer of CattleEye.
“The first insight was deriving its locomotion score, or whether or not the cow is lame.”
He said early identification of lameness enables early intervention and reduces the drag that lame cows have on the overall efficiency of the herd.
“The problem is the lame cow is an inefficient cow. So she’s costing you the same amount of money, producing the same amount of greenhouse gas, but making less milk,” Canning said.
“We’re also looking at body condition scores. We’ve trained the algorithms to be able to recognize these and then give it back to the farmer.”
He said CattleEye helps reduce the carbon footprint of dairy herds by identifying inefficient cattle, and it provides additional value.
“There’s a real requirement for accurate on-farm data, and because we’re not attaching anything to the cows or anything like that, we can find all these people wanting to use our data to be able to look for genomic traits or work out if their nutrition is working,” Canning said.
Downstream from the farm, retailers will be able to show their customers cattle in their supply chains are well taken care of.
Using CattleEye’s algorithms to identify and monitor specific cows helps producers get away from having to attach wearable devices.
Wearable devices “tend not to scale up easily and they require a lot of commissioning. So you find in larger farms they’re not used because you’ve got to go in and attach something to the animal, which is a pain,” Canning said.
He said the cost of the system for small farms is about $35 per cow per year, and the price can drop for bigger herds to as low as $2 per month per cow.
The company is investigating if the system can be deployed in beef herds to monitor disease, and to find the best time to sell an animal.
However, he said the camera works best shooting straight down at the cattle from above, and many beef facilities are not suited for this.
Canning said CattleEye will be available in Canada in the near future.