How to count sheep and stay awake


COUNTING: A project using artificial intelligence and video analysis allows live export companies to accurately count sheep and loading densities in real time.

By Alastair Dowie, Editor of Australian Dairyfarmer

Sheep counting just got easier, and safer, with the development of artificial intelligence and video analysis to track sheep movements in a range of applications.

The University of Technology (UTS), Sydney, project has an application in sheep handling for live exports, saleyards, feedlots and long-haul road transport.

UTS leading chief investigator Dr Jian Zhang said the project, conducted in conjunction with Meat & Livestock Australia and livestock export companies, helped address animal welfare concerns, particularly around sheep and density issues for sea travel journeys of more than two weeks.

Dr Zhang, who will deliver a paper on the project at the Cebit 2019 Expo in Sydney in October, said AI used video analytics for animal welfare and sheep counting in real-time. He said vision and pattern recognition algorithms were able to address the loading environment, including a range of weather and lighting conditions

He said the counting of sheep onto export vessels was a difficult and time consuming process.

Using cameras mounted above the loading races, the program could track the movement of the sheep.

"We can make sure the sheep is always moving forward and not coming back," he said.

Using image segmentation, the program could also calculate the area the sheep needed and whether that area was available on the ship.

"That is the basic condition for animal welfare," he said.

"This technology can tell us whether the ship is over-occupied or not, or whether density is over or not."

Dr Zhang said it also allowed to check sheep and whether they were standing or laying down, in both saleyard situations or on vessels at sea.

"Sheep that have been laying on the floor for long periods, a report can be sent to the supervisor to check the sheep," he said.

Dr Zhang said bullying behavior among the sheep reduced food and water intake of the affected sheep.

This technology monitored the sheep during the trip and animals affected could be identified and a supervisor advised.