Amber Hickman |
The Alexander von Humboldt Institute in Bogotá, Colombia, has partnered with the Microsoft AI for Good Lab to launch Project Guacamaya, which aims to use artificial intelligence to monitor deforestation and biodiversity in the Amazon rainforest.
The organisations will collaborate with Colombia’s CinfonIA Research Centre and the Universidad de los Andes and Instituto Sinchi to use a mix of new AI models for satellite analysis and modified AI models from existing Microsoft projects. Databases are being stored in the cloud, with Microsoft Azure being used to design and train the models.
“If you think about one of the areas in the world that we need to save, it’s the Amazon,” said Juan Lavista Ferres, vice president and chief data scientist at Microsoft’s AI for Good Lab. “This project is not going to solve all the problems the Amazon has, but it is going to solve one I think is fundamental: you can’t solve a problem if you can’t measure it. By doing this, we can actually start measuring the problem.”
The project is combining data from satellites, hidden cameras and acoustic records to accelerate the pace of analysis, so the organisations can gather a better image of the ecosystem and more quickly develop reports for governmental bodies.
The satellites from technology partner Planet Labs are providing daily high-resolution images of Earth, which the AI models can analyse and spotlight areas where illegal mining or deforestation is occurring.
Meanwhile camera boxes have been placed throughout the Colombian Amazon by the project and automatically take photos every time they detect movement. The AI programme can then analyse the photos and alert the researchers to any changes in the ecosystem.
Project Guacamaya is also using bioacoustics to capture sounds from the Amazon and applying an AI model to differentiate between bird and non-bird sounds. This is helping researchers to identify the more than 2,000 bird species in Colombia with over 80 per cent reliability.
“If we can produce models with 90 per cent accuracy instead of 80, we reduce a months-long effort to weeks,” said Pablo Arbeláez, director of the Centre for Research and Formation in AI at the Universidad de los Andes. “Officials can make that information available to the public with a much shorter delay. So, just for that purpose, this information is crucial.”
Microsoft is also collaborating with environmental organisation Imazon and nonprofit PrevisIA in Brazil to prevent problems caused by mining and deforestation by using satellite analysis.