Sound Event Recognition

SERVAL: Sound Event Recognition-based Vigilance for Alerting and Localisation

Sounds of gunshots, chainsaws, vehicles, dogs, or talking people, may signify the presence of poachers or other intruders. Wouldn’t it be great if you could use smart (automated) sensors to hear them? This is exactly what the SERVAL sensor has been designed for.

The SERVAL sensor allows you to extend your vigilant listing-range both in space (more places at the same time) and in time (while you are concentrating on something else). It does so by distinguishing the various sounds in real-time and relaying them to you. Thus giving your the opportunity to intervene in a timely manner.

Its functioning

The SERVAL sensor is a smart sensor, meaning that it can be trained to distinguish all kinds of sounds. To this end we use a combination of known physical sound characteristics and deep learning. The trained neural networks (conv-net) are being deployed on the sensor, which sends out alerts to the SCCSS-platform, from where it can be dispatched to early responders.

The meaning of sound is very context specific. For example, the noise of cars in nature parks is very natural during day time in the tourist season. At night, or outside the touristic season, however these sounds may be very alarming as they may signify the presence of poachers. Exceptions and rules like these are managed in our SCCSS IoT-platform.

Latest news (April 2017)

With the help of Karol Piczak (a befriended Polish researcher) we managed to create a full-circle and real-time sound event recognition system which can be used to send out alerts to e.g. rangers through our fast-response coordination app Cluey.

Relevant sounds for nature conservation which we currently master include gunschots, engines, chainsaws, scooters, motorbikes, cars, trucks, helicopters, footsteps, coughing, sneezing, laughing, talking, breaking, barking, blotting and mooing. All on one sensor. Any other type of sound can be learned and added as well, if we can collect sufficient tagged sound samples of that sound type.

Another very meaningful application is the monitoring of wild elephants. If early warnings can be send to rangers once wild elephants are approaching, they may be able to deter the very dangerous confrontation between wild elephants and villagers. To achieve that, we’ve established formal collaboration with renown wildlife researchers Angela Stoger and Shermin de Silva, with sound recognition researcher Matthias Zeppelzauer, and with M+P, a company specialised in acoustics. With them we are working on a sound-based elephant monitoring and early warning system that may be used to reduce human-elephant conflicts.

Can’t wait to get started? Contact us to discuss how we can help you!