Automating mobility and body condition scoring will help to improve interpretation of cow health and welfare data while saving on labour.
TEXT PHIL EADES
Developments in camera technology and photographic interpretation are opening the door to cost-effective and more frequent recording of parameters, such as mobility and body condition score, as well as providing more timely and consistent information, saving on labour and removing variation in visual assessment.
HerdVision’s Stuart Adams says that regular monitoring activities, such as mobility scoring, are increasingly becoming a requirement of milk processors and assurance schemes. “The challenge is to find the time to carry out herd mobility assessments and then to report the results. The next challenge is doing the assessment consistently,” he explains.
“If two people assess a herd there is a risk they will do so differently. One person’s mobility score 3 could be someone else’s 2. Differences in interpretation could lead to less effective management decisions being made and could delay intervention to address problems.
“These issues can be overcome by using 3D cameras, and the technology used for other valuable herd management measurements such as body condition scoring.”
He says that body condition scoring (BCS) suffers from the same issues of consistency and frequency of assessing cows. As change in BCS is more important than the current score taken in isolation, regular assessment is required to assess trends.
“Ideally, cows should be regularly condition scored from calving until conception, usually around day 90 of lactation. Then they need to be conditioned scored in time to allow them to dry off in target condition, which means scoring cows for the final 60 days of lactation. No dairy business will have the time available to BCS cows this regularly and frequently. “Again, 3D camera technology provides a way to greatly improve the frequency and quality of BCS recording, improving the access to valuable and reliable management data,” says Mr Adams.
Based on an initial concept developed by Kingshay, in consultation with the Bristol Robotics Lab based at the University of the West of England, HerdVision is a static 3D camera system that is now the most widely proven on farms in the UK and Europe.
“The initial work showed the camera technology could identify changes in mobility and condition score. We have now moved the technology to a point where it can work effectively on-farm and provide valuable data,” Mr Adams explains.
“We have worked to ensure cows are correctly identified, because they can push, jostle, mount each other, reverse up and so on. They rarely stand perfectly still under the camera. The system works to constantly improve the accuracy of the results through machine learning, and the HerdVision team cross-references results with a panel of leading independent experts. “Considerable on-farm data comparing the camera against scores from RoMS-accredited scorers show a high degree of correlation, demonstrating the accuracy of our camera.”
Camera system: tech assesses BCS and mobility every time a cow walks under it
The robust and waterproof camera, which is the size of a small shoe box, is mounted above the cattle race, allowing it to take images of individual animals. It can be installed on any race in just a couple of hours and as it is sited high up is safe from physical damage and soiling.
“Accurate cow identification is vital if the system is to work effectively,” stresses Mr Adams. “The technology identifies animals using their EID tag and does not interfere with other EID readers on the farm, avoiding conflicts and giving 100% accurate cow identification.” A 3D video of the cow is produced, with the camera recording 30 frames per second. Mr Adams explains that each pixel on the camera is a measurement point, allowing a detailed image to be created. The camera identifies and measures key anatomical features required to allow both mobility and body condition scoring every time the animal passes under the camera. “By making more measurements, both more accurately and frequently, the quantity of data per cow is far greater than could be achieved by human assessment,” adds Mr Adams. “By viewing every cow every time she goes through the parlour we get at least two images per day with an unequalled degree of consistency. The camera is often sited on the exit from the parlour, helping to oversee the milking herd. But dry cows or heifers can also be monitored by walking them under the camera.”
The images are interpreted on the camera and shared with a cloud-based system. Alerts are sent to an app allowing prompt action, and it is possible to review the three-second video of the animal as required. More detailed analysis using a range of parameters, including days in milk and yield, can be carried out on the HerdVision website.
“For mobility scoring, producers will be alerted to any cow that is showing as lame at the latest pass under the camera. This means they get really timely information, which can be crucial. If a cow can be identified and treated within 24 hours, the severity, treatment costs, and impact on production will be greatly reduced. This is particularly important for cows at score 3, and can have a significant impact on cow welfare,” says Mr Adams.
To look for changes in BCS, every cow is assessed against a rolling seven-day average. “The system highlights meaningful changes, as cattle lose or gain body condition. No other independent system reports on BCS with this level of accuracy and frequency.
“By providing timely and accurate data, the system can play a big role in improving decision making, whether at an individual animal, group or whole herd level. But for many units the biggest benefit will be the saving in labour because the assessments are made automatically and with far greater consistency.
Mr Adams says users are reporting a 20% reduction in the time spent treating lame cows because problems are identified sooner, and a 50% reduction in antibiotic use. “And they point to a 100% reduction in the time taken to record and collate mobility data. More frequent and consistent recording will allow herds to improve health and welfare, which will work through to better profitability.”