While the first phase will focus on deforestation, IIASA plans to use the model for other environmental concerns where crowdsourcing would be applicable.
Deforestation of the Amazon and other forest habitats is a major cause of climate change, as these rainforests are a key generator of oxygen and a sponge to absorb carbon dioxide in the atmosphere.
Tracking deforestation is a difficult process, especially in the Amazon which spans 5.5 million km, roughly the size of Australia. Satellite images are an excellent resource, providing researchers with a tool to check real-time changes, but sifting through thousands of images can be time-consuming.
Enter SAS, which has partnered with the International Institute for Applied System Analysis (IIASA) to build an artificial intelligence model to track deforestation at scale.
“The urgency required to address these transformations requires the application of best of breed technology solutions,” said IIASA Director General, Albert van Jaarsveld.
“By combining the power of our environmental science research platforms, SAS’ AI and computer vision technologies, and the sheer intellectual power of concerned citizens, we will develop AI models that will exponentially increase the value of human insights and strive to deliver near real-time assessment of global environmental change.”
To begin, IIASA will work with volunteers around the world to feed the model with data. This will be done by reviewing satellite images and deciding if the rainforest has been affected by human development. Once the model reaches a high accuracy level, it will take over assessing the images.
“With the power of the crowd and the computer combined, SAS and IIASA will have a robust, extensive platform for empowering citizen scientists to assist in cutting-edge research that helps us all better understand, predict and thrive in the world we share,” said Jaarsveld.
While the first phase will focus on deforestation, IIASA plans to utilize the model for other environmental concerns, where crowdsourcing would be applicable.