Sound Event Recognition-based Vigilance for Alerting and Localisation.
Where humans are, there is noise. Although rowdiness of humans may be irritating, we use it in our favour. Indeed, poachers too make noises that give them away. The challenge is whether we are able to distinguish between ‘normal’ human sounds and those of poachers at work. And, of course, whether we can pick up these sounds in remote and desolate areas.
To determine the meaning of a sound, one has to interpret it. The sound of a car, for example, may be very common in areas where people go on safari, but much less so at night. This is true for most types of sounds. Their meaning varies with its context.
To develop a smart sensor which can interpret contextualised sounds, SensingClues sought the help of M+P (see earlier post) and Sensory Cognition Group of the Rijksuniversiteit Groningen. Both are reputable for their acoustic expertise. Both accepted the challenge.
The working name for our sound sensor is SERVAL, an acronym for Sound Event Recognition-based Vigilance for Alerting and Localisation. Once we finished the design we will search for a funds to develop it, and add it to our anti-poaching toolkit.
We help rangers to turn wild spaces into safe havens!