Analysis of audio recordings for behavior monitoring of zoo animals

Context

In the Karlsruhe Zoological Garden, endangered animal species are specifically bred and thus preserved. In order to keep the animals in a species-appropriate manner, keepers and veterinarians must be familiar with the animals' behavior and adapt environments and situations as necessary. Audio recordings, especially of breathing sounds, can offer conclusions about animal well-being and stress. Accordingly, these will be worked with as part of the final thesis to allow automated evaluation of behavior.

Aims of the work

• Evaluation of the audio signals for statements about the behavior of the zoo animals
• Investigation of the possibilities for statistical evaluation of night activity (e.g. standing/lying sleep in elephants)
• Development of a real-time capable algorithm on a given hardware platform (MA only)

Prerequisites

• Basic knowledge in signal processing
• Interest in machine learning
• MA: Knowledge in hardware related programming (c)