A combination of cloud-based artificial intelligence (AI) and sound detection devices is being used in the fight against illegal logging and deforestation.
According to non-profit organisation Rainforest Connection, almost a fifth (17%) of global carbon emissions come from deforestation.
This is more than all the transport systems in the world combined, and an incredible 90% of all logging in the tropics is carried out illegally.
In the fight to stop this illegal trade, Rainforest Connection has installed devices known as ‘guardians’.
These small sound-detection devices are placed in the forest canopy and stream the sounds they detect to the cloud.
The main problem is that it takes time to analyse the sounds and identify the chainsaw noises that indicate logging activity.
By the time that rangers have reviewed the data and sent a response team, 14 days could have passed.
In this time, large areas of forest could already have been lost, and the loggers may have moved on.
By combining the guardian data with specialist AI analytics, provided by Hitachi Vantara, they hope to massively speed up the process.
AI can separate man-made sounds from the natural sounds of the rainforest
The algorithms create a ‘baseline’ of natural rainforest sounds, making it easier to identify anomalous sounds that are not part of the bio-acoustic signature.
This could include sounds that indicate loggers scouting out a new location, allowing for a response to begin before the chainsaws even start.
The system has so far been trialled in the Sumatran rainforest in Indonesia and is to be rolled out more extensively next year.
Josh Siegel, leader of Visual Co-Creation Services at Hitachi Vantara, said: “We now have up to five days, sometimes more, of a very accurate prediction – a high likelihood that illegal logging will be taking place.
“In our minds it is moving from detection to prediction.”
He added that the system allowed rangers’ response times to be cut by more than a third (35%).
With each guardian covering around three square kilometres, this could cut out around 20 tonnes of Co2 emission, the equivalent of removing 3,000 cars from the roads.
The system is similar to an AI solution being deployed to identify poachers’ gunshots on game reserves, which was developed by the Zoological Society of London (ZSL) in partnership with Google Cloud.
Today’s news was brought to you by TD SYNNEX – the UK’s number one solutions distributor.
