This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Read our privacy policy

Saving the rainforest by listening

HUAWEI CLOUD AI and ModelArts helped RFCx achieve more accurate identification of chainsaw noises and truck sounds to protect the rainforest.

Jul 2018


Topher White is the chief executive officer of Rainforest Connection (RFCx), a non-profit organization that upcycles old cellphones to help protect the rainforest from illegal logging. After seeing Topher’s great work through his inspirational Ted Talk, Huawei reached out to see if they could collaborate and help this incredible cause. It turned out that RFCx were already using a large portion of upcycled Huawei phones due to their impressive reliability at protecting the rainforest, and ability to work nonstop for up to 2 years.Huawei began conversations with RFCx and are now working with them to build a fully connected intelligent ecosystem across multiple rainforests.

Customer Snapshot

Founded in 2014 by ITER physicist and software engineer Topher White, Rainforest Connection has projects in 10 countries on 5 continents, including California, Costa Rica, Ecuador, Peru, Bolivia, Brazil, Romania, Cameroon, South Africa, Sumatra. They are a NGO and are based in San Francisco, California. They help protect the world from illegal deforestation and illegal poaching, and they have already worked with Google and Amazon.


RFCx is working with a complex set of technologies to help protect the rainforest. Not only are they fighting illegal deforestation and poaching, but they are also learning more and more about the animals themselves. RFCx creates solar powered audio monitoring systems called Guardians. These Guardians use old Huawei cellphones as the heart of their system. With help from the Cloud and telco towers, a call is being made every second of the day to AI equipped servers that monitor the dense and complex sounds of the rainforest in pouring rain, blistering sun and dripping humidity. If any illegal sounds are heard, such as chainsaws or trucks , rangers are notified in real-time and sent the location for further investigation. RFCx and Huawei are now using their partnership and ecosystem of AI technology to better understand animal sounds in the rainforest—leading to some much-needed help for several endangered species.

The first challenge is how to collect and transmit sound data in an environment of high temperature, high humidity and no fixed power supply, and store and manage this huge and growing data on the back-end platform safely and efficiently. Second is how to make real-time and rapid analysis of these data and accurately tell the location where the logging was carried out. The sounds in the rainforest are complicated and the data is huge. Therefore, a high-precision recognition algorithm is needed to realize the automatic recognition of the fraud.


Right now Huawei and RFCx are working together to develop an innovative platform that includes equipment collection, storage services, and intelligent analytics. The equipment collection comes from the upgrade of Huawei's old mobile phones. They can work around the clock for two years, collecting sound data and uploading it to the cloud in real-time.And then Huawei's powerful big data service is used to store and manage the audio data collected from various collection points. At the same time, Huawei has begun to cooperate with RFCx to develop a more accurate intelligent algorithm model based on Huawei's advanced artificial intelligence service (HUAWEI CLOUD AI) and tools (ModelArts) to achieve more accurate identification of chainsaw noises and truck sounds. In addition, Huawei is helping RFCx build intelligent models that detect and analyze spider monkeys' sounds, providing information about their habitat, threatened information, and even life habits, helping foresters protect endangered species.

Huawei technicians are testing the collected spider monkey sounds on the ModelArts platform.

“AI basically will allow me to train a machine, to train an algorithm to detect the species. At the moment I am collecting 200,000 of data. It will take 60 years of my life, every single day if I want to listen to it. So, without AI, we can’t physically analyze these large data sets. We will then use that data to find the different calls of the animals, and then use these data to give us a distribution map of where the different animals are.”

Jenna Lawson, PHD researcher, Imperial College London


RFCx are currently using old Huawei phones to protect over 2000 square kilometers of land. RFCx is also using Huawei's Cloud offering to upload data. Furthermore, Huawei’s AI team have completed their first voice detection model, which has been tested successfully and shows a vast improvement in accuracy over RFCX's previous model. Bourhan Yassin, the COO of RFCx, helped explain further how Huawei is helping to improve their technology: “by having a high precision model, which we are hoping to achieve with Huawei’s help, will significantly reduce the false positive rate and eliminate user fatigue related to invalidating false positives.”