Scoring error highlighted in cow condition trial

The considerable margin for error when condition scoring dairy cows has been highlighted in a recent farm trial where over 10% of cows with a herd went uncategorised. 

A total of 184 high-yielding Holstein-Friesian cows were body condition scored on three separate occasions, with human error resulting in a total of 21 cows passing through handling unchecked.

Trialists at dairy technology manufacturer DeLaval and The Dairy Group, who carried out the trial, said cows were either misidentified, incorrectly recorded, or issues were encountered with cow identification or handling.

See also: Automated body condition scoring hits UK farms 

Camera claims

  • Reduce ketosis rates by 50% and improve conception rates by 50%
  • Boost yields by 545kg a cow in first 120 days of lactation
  • Save farmer time and feed costs through automatic scoring and trimming overweight cows

Three separate BCS assessments were undertaken by three different scorers:

  • the farmer
  • consultants from The Dairy Group
  • and DeLaval staff.

They concluded the DeLaval BCS system, a camera that analyses cow shape as they pass underneath, is more accurate than the human eye and requires no farmer time. Average “manual” BCS scoring took three hours compared to zero time for the automated system that required zero input from the farmer. 

“The cows on the trial farm are scored once a month by the farmer himself,” Kieran Fitzgerald of VMS Solutions told Farmers Weekly. “They are keen on BCS scoring at the farm and it’s a key part of management.”

He added that most farmers now understand the benefits of BCS but fail to have the time and maintain a consistent score.

Mr Fitzgerald said: “The technology is a world first and completely changes the way the farming community will look at body condition scoring. Achieving an accurate body condition score is now extremely simple. It gives consistent and daily feedback on the condition of a herd without any work on the part of the farmer.”

Condition camera

Launched last autumn, the camera uses 3D imagery and an algorithm to create a snapshot of a cow’s condition.

BCS Camera

  1. Cow passes under camera
  2. Camera recognises cow
  3. Camera selects best still image in video sequence
  4. 3D image converted into a body condition score

See also: Benefits of using an automatic body condition scoring camera

Algorithms can currently only assess four breeds: Holstein, Norwegian Red, Swedish Red and Fleckvieh, expected to increase to six by the end of the year when Jersey and Jersey-cross (Kiwi-cross) are finalised.

There are currently six cameras in use in the UK. Two conventional dairies in Scotland are using the camera and three robotic (voluntary milking systems), as well as the trial farm in the South East of England.

Harper Adams University has plans to install a system this year.