Chicken CCTV: Tech solution for poultry problems

A low-cost camera set-up could net major savings for Australia’s multi-billion dollar chicken industry.

University of Southern Queensland (USQ) mechatronic engineer Dr Cheryl McCarthy is researching a new way to detect chicken welfare using machine vision, a camera technology with artificial intelligence akin to face recognition.

Chicken sheds are typically inspected three times per day and rely on the human eye, whereas Dr McCarthy’s system deploys continual on-board image analysis to provide minute-by-minute alerts.

“The cameras observe the daily behavioural patterns of the chickens, picking up changes indicative of underlying health issues,” Dr McCarthy said.

“The camera technology's artificial intelligence spots subtle patterns in behaviours – are the chickens sitting more often? Grouping together? Not eating as much?

“This data can be used in the diagnosis of temperature stress or foot conditions like footpad dermatitis.”

The software has wide-reaching possibilities, with investigations also underway to apply it to chicken weight estimation and in a free-range setting using drone technology.

“The benefits include increased productivity via early interventions and better animal welfare, and as the cameras are low-cost, the technology is well-suited to on-farm use,” Dr McCarthy said.

The research is the first of its kind in Australia, and has been the subject of a two year project funded by AgriFutures™ Chicken Meat Program.

It follows Dr McCarthy’s previous research in remote monitoring using smart cameras to perform tasks like cattle condition scoring, weed spot spraying and crop growth monitoring.

Key facts:
• The Australian chicken meat industry is worth $2.87 billion (gross value of production 2018-19), with an approximate retail value of around $6.6 billion.
• The industry has a strong presence in rural and regional communities, directly employing about 58,000 people (full time equivalent) across the country.

Four baby chickens