Rumination is closely linked to animal health and physiology, and the ability to accurately and remotely detect this behaviour will be a valuable management tool for enhancing animal health, according to CQUniversity’s Dr Anita Chang.
Dr Chang’s findings mean that critical events like calving, oestrus and the disease status of an animal could all be identified without physically going to the paddock, offering producers not only labour savings but an enhanced level of animal welfare.
“In the long term this could be a game-changer for the way cattle producers monitor and manage their herds, once smart tag devices become more readily available and incorporate algorithms such as our rumination detection measure,” she said.
In Dr Chang’s trial accelerometer ear tags were fitted to eight cows to measure rumination behaviour by detecting changes to their movement patterns and used a variety of different machine learning algorithms to test for the ability to predict rumination.
The results have now been published in Computers and Electronics in Agriculture.
“This paper has a specific consideration for commercial applicability and was focused on achieving industry outcomes,” Dr Chang said.
Two methods were used to detect rumination.
The key difference in the two methods was the data that was used. The first method achieved 98.4 per cent accuracy using individual animal data to create a ‘customised’ model for each animal, while the second method, which is more applicable to a commercial setting, used herd data to achieve an accuracy of 86.2 per cent.
Dr Chang has already been approached by commercial sensor companies to incorporate her modelling into the systems they are developing for use by farmers.
“There has been some work done previously using accelerometers to detect rumination in cattle, however the accuracy was significantly lower than the methods employed in CQUniversity’s approach and didn’t take commercial settings into account,” Dr Chang said.
“In our trial ear tag accelerometers were used, due to their easy application and commercial relevance, rather than accelerometers attached to collars which were used in most previous studies.
“On-animal sensors are revolutionising livestock farming ... the ability to remotely monitor animals through on-animal sensors offers huge labour savings and productivity gains, particularly for producers operating extensive livestock systems.”
20°C