What do media technology, artificial intelligence and nature conservation have to do with each other? A lot, as William Menz proved in his master's thesis. He wrote it in the AVT group at the Institute for Media Technology at the Technical University of Ilmenau and thus very successfully completed his studies in media technology.
For many nature conservation projects, it is important to know the size of animal populations and to monitor their changes. This is often done with photo traps, because this way the animals are not stressed and the populations are not influenced.
In the master thesis of William Menz, which was done in cooperation with the company Eurofins MITOX, the focus was on monitoring the mouse population near intensively farmed areas. For this purpose, Mr. Menz developed photo traps equipped with decoy food and took tens of thousands of photographs with them. The previously known problem was that there are several species of mice and other small mammals that frequent the traps. For accurate monitoring, these species must be distinguished. For this purpose, artificial intelligence methods can be used to relieve the biologists of the tedious work of evaluating the photos. Mr. Menz tested different algorithms, selected the appropriate ones and trained the neural networks with a few thousand selected photos. As a result of the work, large amounts of image material can now be automatically evaluated and the animals classified. This enables regular monitoring of populations in critical areas.