Automated analysis of gunshot records could to make the fight for the preservation of species more effective threatened in the world.
The Zoological Society of London (ZSL) has a new tool in the fight against the poaching of protected species. In the Dja Faunal Reserve in Cameroon, it has placed acoustic sensors capable of continuously recording events up to 1 km away. Based on Google's artificial intelligence (AI) techniques, the device automatically filters out shots fired and thus alerts anti-poaching patrols. Incidents can be pinpointed on a map of the reserve to help locate areas popular with poachers.
It's about planning patrol deployment areas at the time of day when you most expect illegal activity - Anthony Dancer, Director of Technology at the Zoological Society of London.
To date, wildlife conservationists have relied on photographic traps to track poachers. The traps are activated when there is movement and limited to a short distance. Acoustic sensors not only have the advantage of detecting incidents farther away and within a 360-degree radius, but they are also less expensive. At around 70 Fr. per sensor, however, this is still an investment for the zoological society, which has to raise funds.
Anthony Dancer, Director of Technology for ZSL, hopes that this sound tracking of the shots will only be a first step. He envisions
to be able to directly identify the poachers' voices. The sensors could also be used to observe how the wildlife is doing. Sound gunshot recognition technology has already been used by American police. But its reliability is not yet sufficiently established. "One of the most difficult things about gunshot identification is that sound bounces off surfaces," Professor Mark Plumbley of the English University of Surrey told the BBC. In cities, these are usually buildings, cars or people. So technology can be inaccurate." He believes that large, unpolluted desert areas are good for this technology, but there's always vegetation or other things that can disturb it.