Monitoring individual cow feed intake is important for improving dairy farm efficiency. The cost, time and needed maintenance of off-the-shelf systems make them impractical for commercial farmers. We developed a system for measurement and estimation of the feed intake of all cows in a herd during all the lactation period with the goal to fit farmers' requirements. The feed intake was measured by a system consisting of four principal parts: (a) a hanging weighing system (simplified scales hanging on a single load cell), (b) a visual cow identification, (c) an automatic cleaning system, and (d) an algorithm for feed intake estimation based on a linear mixed effect model including milk yield and content, body mass, meal duration and frequency. The system was validated during a two-month experiment with six scales and 12 cows in a barn modelling a commercial one. The validation experiment showed that the system fulfilled the requirement: the three most inefficient cows were found by feed mass measuring. In an example of using the feed behaviour model two inefficient cows were correctly predicted. The scales were accurate within 50 g; the visual cow identification rate was greater than 96% and routine farm practices continued as usual, though with delay. The average cost for a feeding station was about $1,500. Thus, the system can potentially be used for ranking cows by their efficiency in commercial facilities. © Precision Livestock Farming 2019 - Papers Presented at the 9th European Conference on Precision Livestock Farming, ECPLF 2019. All rights reserved.
Monitoring individual cow feed intake is important for improving dairy farm efficiency. The cost, time and needed maintenance of off-the-shelf systems make them impractical for commercial farmers. We developed a system for measurement and estimation of the feed intake of all cows in a herd during all the lactation period with the goal to fit farmers' requirements. The feed intake was measured by a system consisting of four principal parts: (a) a hanging weighing system (simplified scales hanging on a single load cell), (b) a visual cow identification, (c) an automatic cleaning system, and (d) an algorithm for feed intake estimation based on a linear mixed effect model including milk yield and content, body mass, meal duration and frequency. The system was validated during a two-month experiment with six scales and 12 cows in a barn modelling a commercial one. The validation experiment showed that the system fulfilled the requirement: the three most inefficient cows were found by feed mass measuring. In an example of using the feed behaviour model two inefficient cows were correctly predicted. The scales were accurate within 50 g; the visual cow identification rate was greater than 96% and routine farm practices continued as usual, though with delay. The average cost for a feeding station was about $1,500. Thus, the system can potentially be used for ranking cows by their efficiency in commercial facilities. © Precision Livestock Farming 2019 - Papers Presented at the 9th European Conference on Precision Livestock Farming, ECPLF 2019. All rights reserved.