A SENSOR-BASED system to give real-time estimates of sheep liveweight, fleece production and disease status is being developed by the Commonwealth Scientific and Industrial Research Organisation with the aim of going beyond current in-paddock weighing systems.

CSIRO said walk-over weighing systems can often be labour-intensive, rely on voluntary animal use or provide measurements that include the weight of fleece.

The research agency said it is combining expertise in 3D modelling and reconstruction with livestock science to develop a multimodal sensor system that produces a high-resolution three-dimensional image of each animal as it moves in view of the sensors. Using artificial intelligence, the proposed system would calculate accurate estimates of volume, liveweight, and fleece weight from the images it captures. Each measurement can then be linked to an animal’s electronic identification (EID) tag.

The new technology enables regular, non-invasive monitoring throughout the growth period, supporting efficient, cost-effective and sustainable management decisions including early disease detection, the agency said.

CSIRO said the technology also has potential applications beyond sheep, with opportunities to expand into cattle and other livestock industries in the future.

The research agency is seeking producers and industry stakeholders interested in engaging with the research team at key points throughout the project to help ensure the system is fit-for-purpose and scalable.

CSIRO principal research scientist Dr Sabrina Greenwood said the system addresses a key gap in on-farm tools for data-driven decision making.

“There are some great technologies coming to market that are helping us make data-driven decisions on-farm, but producers still lack quick, remote ways to accurately measure key animal traits.

“Our system fills that gap by delivering accurate estimates of sheep liveweight and fleece weight, without slowing farm operations or requiring animal contact,” Dr Greenwood said.

CSIRO principal technical program manager and FlockMate project leader, Lian Flick, said the tool marks an exciting shift in bringing 3D computer vision research out of the lab and into real-world agricultural settings.

“With recent advances in AI and computing that no longer depend on large servers, we can now process complex data in real time out in the field,” Lian said.

“With these advancements, this system offers the Australian sheep industry information it hasn’t previously had.”

Producers and others wanting to get involved in the project should contact, Lian Flick at [email protected]

The project has received co-investment from Meat & Livestock Australia and Australian Wool Innovation.