SAS recognised for AI-powered deforestation detection app

This is one of many social innovation projects from SAS employees, whose curiosity inspires them to use data to solve pressing humanitarian issues.

SAS was recognised in the AI and Data category of the Fast Company World Changing Ideas Awards, which honour the businesses, policies, projects and concepts that are actively engaged and deeply committed to pursuing innovation when it comes to solving health and climate crises, social injustice or economic inequality.

Similarly, SAS was recognised in the Environmental Stewardship category of the PRNEWS CSR & Diversity Awards, which honour communicators who use their platforms for the betterment of their communities and the global community at large. 

Detecting deforestation, one click at a time

The Amazon is the world’s largest rainforest and home to an impressive 10% of the world’s biodiversity. Unfortunately, deforestation is wreaking havoc on the Amazon; approximately 800 square kilometres of forest is destroyed every month. Even worse, during the COVID-19 pandemic, there’s been a surge of deforestation in Brazil. 

To drive stronger deforestation policy responses, SAS and IIASA jointly built an AI platform to analyse satellite images of the rainforest to show the location and magnitude of forest damage more effectively. Then, they asked an army of data scientists to review the images to better train the AI models.

Users classified if there were signs of human impact in the images, helping improve the AI algorithm and expedite the analysation process. This improved view of deforestation helps support critical policies to quickly and effectively protect our forests. 

The project launched on Earth Day 2020, and throughout the year, researchers continued to add more images from across the Amazon to expand the effort. To date, citizen scientists from across 119 countries classified more than 919,000 square kilometres of the Amazon. By using image data from this ecologically diverse territory, the computer vision model is given a variety of examples so it can eventually learn to detect human impact anywhere in the Amazon.

In addition, citizen data scientists have started classifying more recent images to identify where new human impact is occurring. If researchers can show their predictive model is successful at identifying areas most at-risk of future deforestation, it could be useful for governments and forest monitoring bodies to carefully track and respond to forest changes.

SAS is continuing work to solve this important issue and has recently joined forces with Amazon Conservation to expand efforts for developing AI algorithms to expedite the identifying, tracking, and intervention of illegal deforestation in key parts of the Amazon.

Learn more about innovation at SAS and its efforts to use data to drive positive global change as part of the Data for Good movement.