Will Marshall on combining satellite data and AI for sustainability
At a recent Stanford Doerr School of Sustainability Dean’s Lecture Series event, the Planet Labs CEO discussed how the marriage of satellite imagery and artificial intelligence creates new opportunities to track renewable energy projects, wildfire damage, deforestation, coral reef systems, and more.
As he took the microphone at a recent Stanford Doerr School of Sustainability Dean’s Lecture, guest Will Marshall, CEO of satellite data provider Planet Labs, did not waste time: “I think there is a profound underutilization of the new data sets that we have from satellites in how they can have a real impact in enabling us to transition to a sustainable economy and a sustainable planet,” he said.
Marshall had been introduced by the day’s host and moderator, Dean Arun Majumdar, with a nod to possibilities opened up by the marriage of satellite imagery and artificial intelligence: “We are living in the age of AI where data becomes very, very important. But to really lean into this, you need partnerships,” Majumdar said.
Above and beyond
Marshall, who got his start working at NASA designing lunar orbiters and nanosatellites, co-founded Planet Labs in 2010. The company now designs, builds, and operates a fleet of more than 200 imaging satellites that provides a daily flood of geospatial data. “Planet Labs started planning about 12 years ago for a goal of imaging our spaceship – spaceship Earth – in its entirety, every day,” Marshall said.
Each Planet Labs satellite captures a strip of images of the land below, and each pixel captures data from a patch of Earth just 3 meters by 3 meters. The satellites take photographic images in the visible spectrum, as well as in eight spectral bands beyond the range of human vision – infrared, ultraviolet, and more.
“We can revisit any place on Earth up to about 10 times per day,” Marshall said. “Twenty-four-hundred layers of imagery for every point on Earth’s 150 million square kilometers of land mass.
Advanced vision
All this data is almost too much for humans to digest, but with the advent of advanced computer vision – computers able to see and interpret images – artificial intelligence is helping to make sense of it all.
Analyses of geospatial data using AI can help researchers monitor natural disasters like floods, droughts, earthquakes, and tsunamis, and track reforestation or the expansion of renewable energy installations. Working with Microsoft, Planet Labs recently mapped solar and wind energy facilities around the world, Marshall said.
Planet Labs satellites can look at neighborhoods ravaged by fire and cities ripped apart by war. AI enables assessments of building damages that can then inform estimates of rebuilding costs, as the company is doing in Lahaina, Hawaii, and Ukraine. These data have helped to inform decisions by the American Red Cross in Lahaina, Marshall said. Planet Labs satellites are also scanning the world’s forests for new road starts that can provide early indicators of illegal deforestation, narcotics trafficking, and mining.
Recently, Planet Labs moved beyond land-based analysis to help map coral reef systems across the entire world’s tropical zones, completing a census of coral species making up those important reefs. “Even I was surprised that we could do that,” Marshall said.
Looking ahead
Soon, with the help of AI, Marshall said he can envision a “planetGPT” in which researchers query global information systems (GIS) databases simply by typing questions in a box, no coding necessary. AI would churn through the data and tabulate an answer in minutes. Planet Labs satellites are sending down 4 million images every day, Marshall said. It is too many to look at in person. “The only way to get value out of this is artificial intelligence,” Marshall said.
It all makes for an exciting future for sustainability-minded researchers, Majumdar said, kicking off a vibrant question-and-answer session that closed the event: “This is an amazing time to be a graduate student or even an undergraduate student.”
Explore More
-
Majumdar discusses academic freedom and the resources that support research: "partnerships are based on trust."
-
Phasing out animal agriculture represents “our best and most immediate chance to reverse the trajectory of climate change,” according to a new model developed by scientists from Stanford and the University of California, Berkeley.
-
Stanford University scholars discuss the Biden administration’s early actions on environmental issues in the Arctic and how the U.S. government can address threats to ecosystems, people and infrastructure in the fastest-warming place on Earth.
