Cell phones in Amazon trees alert rangers to illegal logging, record wildlife

Rainforest Connection is now leveraging machine learning to save rainforests

It’s been just over a year since we introduced you to an ingenious application which makes use of solar-powered cell phones placed high in jungle canopies to help park rangers — and regular citizens — join the fight in saving Amazonian rainforests. Dubbed the Rainforest Connection, the app uses phones hidden in trees throughout the jungle to detect sounds associated with illegal logging.

When the phones pick up these sounds, they alert local park rangers who have the app installed on their own smartphones, giving them the ability to react accordingly. The project, begun in 2014, has now grown up quite a bit, and is now applying machine learning in order to further improve its capabilities.

The app is the brainchild of Topher White, a renowned physicist and engineer who became involved in rainforest conservation in 2012 after volunteering in Borneo with a gibbon protection program. At its conception, Rainforest Connection only listened for certain harmonics, but with the introduction of machine learning, its capabilities have grown enormously.

“Now we’re talking about detecting species, gunshots, voices, things that are more subtle,” he told TechCrunch in an interview. “And these models can improve over time. We can go back into years of recordings to figure out what patterns we can pull out of this. We’re turning this into a big data problem.”

The new machine learning capabilities are a result of White’s work with Google’s TensorFlow platform, and are especially adept at identifying patterns that might otherwise go undetected. As an example, White said, “We should be able to detect animals that don’t make sounds. Jaguars might not always be vocalizing, but the animals around them are, birds and things.” Listening stations can be situated up to 15 miles away from a cell tower.

Rainforest Connection could revolutionize conservancy efforts across the world. While White was in Borneo, he discovered that local conservationists spent nearly half their budget trying to fight off illegal loggers. The problem was that most of their vigilance relied on satellites to detect illegal activities, and by the time park rangers arrived, it was entirely too late.

“If you pick it up on a satellite, not only is the damage already done but at that point, the stakes are so much higher,” White told Digital Trends. “The trees have been cut and they’ve carried out trucks full of logs.”

This means the logging operation invested time and money into its work and feels inclined to fight to keep what they have. On the ranger’s side, a crime’s now been committed, and there’s a greater drive to prosecute or seek other means of affecting consequences. What this leads to are standoffs and other undesirable scenarios.

According to White, somewhere between 50 and 90 percent of all logging performed worldwide is illegal

“If you can actually stop them on their way in by detecting a vehicle or detecting a chainsaw before a lot of damage has been done, the stakes are so much lower for both sides that you can actually talk it down,” White said. “In central Africa and throughout Latin America, both sides are pretty well-armed. The quickness of the alerts and the quickness of the reaction is super important for that same reason which is that you don’t want the stakes to be high. You want it to be much simpler to just turn the truck around and leave than to actually have a showdown.”

According to White, somewhere between 50 and 90 percent of all logging performed worldwide is illegal. That said, recent research found that if rangers show up and stop poachers once or twice, they tend to leave — until the next logging season, at least. In that way, the app remains highly effective at reducing clandestine operations. The units which go up in the trees, called guardian phones, are recycled cells which are typically five to six years old — however, they’re still “powerful machines,” according to White.

solar powered cell phones saving amazon rainforests rainforest connection 3

Powered by the sun and operating 24 hours a day, the units record all sounds in the forest. Once recorded, the phones process each sound, crunch the data, send it to the cloud, and then analyze it in real time. White outfit the analytics to detect chainsaws, logging trucks, vehicles, gunshots, certain species of animal, and other relevant noises. The RFCx app received a huge boost when outdoor sports giant Mountain Hardware took it under its wing as part of the sports company’s Impact Initiative, which supports various eco-campaigns.

Along with deterring illegal logging, the app also offers a platform for ecologists to study wildlife. Soon, they will be able to access up to 18 months of data across hundreds of locations.

“If you’re a scientist who wants to study a certain bird in the Amazon, in the past you used to have to get a grant and organize it and go down there for a few months and record (everything),” White added. “There’s so much we can discover about the forest without needing people to actually go there.”

“There’s so much we can discover about the forest without needing people to actually go there.”

Furthermore, anyone has the ability to use the app to listen to the sounds of the rainforest in real time and get alerts about what’s happening. Users simply open the app on a smart device to hear live birds, insects, monkeys, and other natural sounds from wherever they are in the world. This, in itself, adds momentum to conservation efforts, White pointed out, by getting average citizens interested in what’s going on. Being able to listen to those sounds and viscerally connecting with the forest increases people’s sense of personal investment. Lots of citizens would like to get involved, he said, but don’t know how. The app lowers the bar for what it takes to get involved.

“The bar is downloading an app onto your phone,” he acknowledged. “That’s all you have to do. You don’t necessarily have to pay anything, and we’ll send you alerts when something happens and you can learn about the forest that way. That’s all it takes for you to actually make a difference.”

According to White, in the context of a rainforest, audio is actually preferable to video feeds.

“It’s experiential and imaginative in a certain way,” he said. “Video is kind of what we’re used to, but the truth is that video in a rainforest isn’t nearly as interesting as audio. You’re just going to see a bunch of leaves in front of you unless the camera picks up some leopard, which is pretty rare. But [with audio] you can hear all sorts of animals that are out there. You don’t even know what most of them are — we don’t know what most of them are. It sounds like a Star Wars laser battle.”

White is now looking for ways to involve other people in his project. The developer has already partnered with Google to kick off a pilot program that allows kids to build their own phone kits.

“We’ve already done three schools and I think a couple hundred students, plus three more in about half an hour,” he told TechCrunch, “And all these devices will be deployed in the Amazon over the next three weeks. On Earth Day they’ll be able to see them, and download the app to stream the sounds. It’s to show these kids that what they do can have an immediate effect.”

He concluded, “An important part is making it inclusive, proving these things can be built by anyone in the world, and showing how anyone can access the data and do something cool with it. You don’t need to be a data scientist to do it.”

Updated on March 26: Added information about Rainforest Connection’s new machine learning capabilities.