Wildlife Disease Warning System WWS – AI-supported camera-based system solution for the early detection of wildlife diseases

Diseases can nowadays be transmitted very quickly between wildlife, livestock and humans. The sharp rise in livestock numbers, together with the increasing mobility of animals and humans, raises the likelihood of disease transmission between wildlife, livestock and humans. This makes the rapid spread of pathogens increasingly likely and threatens wildlife and livestock populations as well as human health. Early detection of animal diseases such as African swine fever in wildlife populations is therefore highly relevant economically, ecologically and also for our own health.

The aim of the WWS project is to develop an early warning system for the use case example of the African swine fever that is able to detect animal diseases using camera sensors and is therefore non-invasive. To this end, disease-related changes in the movement patterns of infected animals are to be recorded and analysed with the help of artificial intelligence (AI). The advantage of this is that diseases can be detected in living stock without the need to capture and sedate an animal in order to take blood samples, which also take several days to analyse in the laboratory. The health status of individual animals allows precise conclusions to be drawn about the outbreak of a wildlife disease in the population concerned, which can thus be detected at an early stage. If a disease outbreak is recognised at an early stage, it can be combated much more efficiently and effectively – the spread of the disease can thus be contained and the ecological, economic and health consequences significantly mitigated. The ‘early warning system’ to be developed as part of this collaborative project therefore offers authorities in particular a highly efficient tool for quickly identifying wildlife diseases and intervening accordingly at an early stage.

In the WWS project, the Leibniz-IZW is developing a set of parameters for characterising wildlife disease effects. It should be possible to record these parameters with camera sensors so that they can be utilized for the development of an AI-based video analysis. To this end, typical behaviour and movement patterns of animals such as wild boar will be identified and analysed to see how they change in the context of a disease. This could be changes in motor sequences or a slowing down of typical movements. An AI is being trained to automatically perform this classification on video recordings that are captured by autonomous camera systems developed in the project. A cloud-based data communication concept allows the information generated to be transmitted and an alarm system to be set up for the authorities.