Acoustic detection
trackIT Systems is working on new automatic and digital methods to improve the detection of wild animals worthy of protection. Large parts of terrestrial vertebrates such as birds, amphibians or mammals, but also some insect species (grasshoppers, cicadas) make characteristic sounds that can be used to recognize the species. This characteristic is used for the observation and research of corresponding species, in which experts recognize and locate the animals in the field. Automatic audio recorders are becoming increasingly popular for recording bats in particular, but also for birds, producing huge amounts of data, much of which has to be examined manually afterwards.
As we need to obtain reliable data quickly and over a large area in the rapidly advancing biodiversity crisis, this approach is not sensible, as the subsequent analysis of the data incurs high costs in terms of time and money and the availability of experts trained for this task is limited.
Recording and AI-based evaluation of animal sounds
The acoustic detection of wild animals
trackIT Systems, funded by the Distr@l program of the Digital Ministry of Hesse, is developing AI-supported recognition of vocalizations of various animal species on autonomous sensors in the field.
Solar-powered sensor units consisting of a mini-computer (Raspberry Pi), the microphone and power unit and an LTE module are installed at any location.


BirdEdge
Principle of real-time recognition of animal voices
The stations run easy-to-use software that contains neural networks (AI) for recognizing animal sounds (4). The network is optimized in size and architecture for use on minicomputers and enables the classification of recorded animal sounds in real time with relatively low computing power requirements.
Classification is performed on the live audio stream. This means that sound files do not have to be stored and analyzed after recording. This approach enables significant savings in terms of recorded and stored data, as it is possible to define on the devices the species and number of reference recordings to be stored per day (5). For example, not every robin or pipistrelle detected needs to be saved as a recording file.
However, if a very rare species is detected, it should be possible to have this recording checked by experts.
- Minimal workload for recording
- Evaluation of acoustic events at a location in real time
- Suitable for large-scale monitoring programs at hundreds of locations
audio-tracker:audio - Real-time recognition of animal sounds
Support for field work
E.g. site selection for net catches
In use
Transmission and visualization in real time
The classifications are transmitted in real time to a remote server, where they are stored in a database with a time stamp, recognized animal species and confidence level (how certain the neural network was during recognition) and immediately displayed visually.
Expandable with sound dummy
For the detection of particularly rare bird species, it may be necessary to lure them with the help of a sound dummy. The audio tracker can also be equipped with a loudspeaker so that recordings can be played at defined times and for a defined period of time. The playback events and the reactions to them can be viewed in real time from home.

In figures
Our experience
The trackIT system has been developed and tested by the founders over the last 6 years, initially at the University of Marburg as part of the Nature 4.0 project and since 2023 in the company they founded. Since 2019, the founders have been accompanying its use in intervention procedures such as road construction or wind power processes and in research projects. Our customers and project partners range from planning offices to nature conservation associations (NABU, LBV) and authorities (NLWKN, HLNUG).
Acoustic recording in our projects
trackIT Systems’ technology can enable continuous monitoring and detection of wildlife in conservation projects. Our solutions offer automated radio telemetry and AI-powered methods for acoustic detection of birds, bats and amphibians to provide accurate data for infrastructure and conservation projects.

Nature conservation projects
To the projects
Intervention procedure
To the projects