Automated Animal Welfare and Behavioral Analysis in Small Groups of Pigs

Initial Situation and Objectives

In pig farming with small group housing, continuous observation of feeding and drinking behavior, group interactions, and general animal welfare indicators is of central importance. Changes in activity patterns can provide early indications of disease, injuries, contamination, or shortages of feed or water.

Currently, the assessment of these parameters is mainly carried out manually and on an ad hoc basis. This process is time-consuming, documentation-intensive, and heavily dependent on individual experience. A systematic, objective, and continuous evaluation of group-related condition data is therefore only partially established under practical farming conditions.

The aim of the project is to develop a camera-based system for the automated recording and analysis of key behavioral and condition parameters in small groups of pigs.

© Fraunhofer IGD
2D image-based depth estimation in a pig pen.

Technological Approach

The project relies on RGB-based camera monitoring in barn environments, combined with AI-supported image analysis.

The system includes:

  • Computer vision methods such as segmentation and object detection to identify individual animals and resource areas
  • AI models for analyzing feeding and drinking activities, movement patterns, and group behavior
  • derivation of objective parameters related to growth, activity levels, and abnormalities
  • alert and escalation mechanisms when deviations are detected
  • options for remote monitoring and targeted intervention

The solution is designed for robust operation under real barn conditions and enables continuous data collection without additional stress for the animals.

Benefits, Application, and Exploitation

The system supports automated management processes in pig farming through:

  • early detection of signs of disease or deficiencies in supply
  • objective monitoring of animal welfare at the group level
  • data-driven forecasts and trend analyses
  • reduction of documentation workload
  • improved working conditions through remote access and targeted intervention

The solution is designed as an extension of existing barn and management systems and addresses agricultural farms as well as providers of automated livestock management technologies.

Practical Environment

Development and validation are carried out together with consortium partners within the framework of the joint research project AI Animal Welfare

Activity monitoring in the pig pen. Data collection and analysis carried out in collaboration with the consortium partners of the joint project AI Animal Welfare

Funded by:

More about our research into smart farming

 

Back to overview

Learn the essential facts about our team and its work.

 

Livestock

We offer customized solutions for a variety of animal species in the livestock industry.

 

Crop

We combine the advantages of drones and visual computing to perform vegetation analysis and species recognition from the air.