Satellites plus AI: How Planet Labs tracks change on Earth

In This Article:

From potential deforestation efforts to preventing tree branches from taking out power lines, there's a lot Planet Labs' satellites can spy from space. But it's too many images for one person to track. That's where AI comes in.

Planet Labs CEO Will Marshall (PL) tells Yahoo Finance how his company uses images to train its AI algorithms. "We've got this corpus of imagery... to train our algorithms so that you can do things like find any object around the Earth and backwards through time," Marshall explains. Marshall uses the company's project in Brazil, where Planet Labs' images and algorithms are used to detect new roads being built in the rainforest, a sign that there could be deforestation efforts to come.

Watch the video above to hear Marshall explain how companies will use Planet Labs' technology to track their carbon footprint.

For more expert insight and the latest market action, click here to watch this full episode of Yahoo Finance Live.

Editor's note: This article was written by Stephanie Mikulich.

Video Transcript

- When you talk about some of the work that you are doing, you build and manage a fleet of more than 200-- just around 200 satellites that are capturing images here on Earth, daily global change on Earth. When you talk about-- there's been so much hype around surrounding AI, how it is disrupting, how is revolutionizing. Across your industry, how do you see the advancements in AI impacting what you do?

WILL MARSHALL: Yeah, absolutely. Well, a huge amount. So AI is all about the data. Algorithms on their own are not worth anything. It's all about what data you train them on as to how useful they are to you. If you train them on the text of the internet, you can answer questions that you can find on the internet.

We've got this corpus of imagery, 2,500 images for every point on the Earth's landmass on average. So to train our algorithms so that-- you can do things like find any object around the Earth across and backwards through time. Let me give it a very specific example though to make it concrete. We work with the government of Brazil to track deforestation.

Now we produce millions of images all the time of Brazil. No one could look at all those images by hand and check if the difference between today and yesterday, if there's deforestation. So we have an automatic algorithm that's trained using machine learning that looks at all of the Brazilian Amazon and finds any new roads. If it finds a road, it's an early sign of people doing deforestation.

We then send them alerts every day of all the locations across the entire Brazilian Amazon for deforestation events. Last year alone, they did 3,000 interventions to go and arrest people or take equipment. They confiscated over $2 billion worth of equipment. And most importantly, they reduced deforestation rates by 55% in Brazil, leveraging our data.