KAMI - Artificial Intelligence for animal welfare

Photo: Lietze/ATB

KAMI: Artificial intelligence for measuring respiration in dairy cows


If dairy cows are stressed, this has a strong effect on their respiratory rate. Accelerated breathing indicates stress such as heat stress, fear and pain at an early stage and provides information about pathological processes and the severity of the stress. The usual method is counting the number of breaths from the movement of the cattle's flanks. However, visual counting is labour-intensive, time-consuming and requires the presence of a person in the barn, which further disturbs the animals. Previous sensor technologies for recording the breathing rate have to be attached to the animal and often require a WLAN or internet connection in the barn.

With KAMI, we want to create a solution for the first time with which we can record the individual breathing rate of dairy cows without contact and under practical conditions. To do this, we use infrared thermography and depth cameras as imaging methods and process the images with the help of artificial intelligence.


Jul. 2021 - Jul. 2024



An early warning system for the benefit of dairy cow and farmer

The aim of KAMI is to develop an early warning system that alerts farmers when their dairy cows fall ill or are exposed to incipient stress. In future, this could increase animal welfare and avoid financial losses.

The vision of the project: A low-cost, practical prototype as an AI-based system that recognises animals individually and records their breathing using imaging techniques. The new system is to be integrated into existing control and knowledge systems of dairy farms. The animals will be monitored automatically and information on breathing will be intelligently linked with various parameters such as milk quality, movement, eating and chewing behaviour.

Work steps

  • Generation of meaningful image data (thermal images of the nasal area as well as images of the flank area) to detect the respiratory rate of dairy cows
  • Development of an AI model for the automatic detection of the breathing frequency
  • Integration of the recording system into a barn area