KfV – Method for assessing wildlife accident risk on roads

The aim of the project is to develop a method for assessing the risk of a wildlife accident on roads in two model communities (in Upper Austria and Lower Austria) with forest areas and roads crossing them. This can then be used to take targeted protective measures at these locations.

To do this, aerial drones equipped with cameras and thermal imaging sensors will be used to capture footage of the model region to count wildlife. These video recordings will be analyzed using machine learning methods to obtain estimates of the wildlife population in each area. These counts will be repeated after a few weeks each time, both to obtain reliable estimates and to detect changes over time.

From the distributions of the wildlife populations, a model is developed for which roads are at particular risk for wildlife accidents.

The research group AIST is supporting colleagues of the partner research group, the Media Interaction Lab, in this effort.