Poultry farmers spend a great deal of time checking the health and welfare of their chickens. This can mean walking through the flock up to three times a day monitoring for signs of poor health. This project, building on previous work will develop a proof of concept machine vision system with image analysis for early detection of abnormal flock appearance and behaviour. Automated cameras set up at pop-holes observe the daily behavioural patterns of the chickens, picking up changes that could be indicative of underlying health issues. The algorithms built into the camera technology recognise subtle changes in patterns of behaviour and this information then generates alerts for the farmer about flock appearance and behaviour that potentially warrant follow up veterinary inspection.