Posted on May 5th, 2015

GTM – Debunking Satellite Imagery Myths, Part I

As the saying goes, you cannot believe everything you read. And this adage could not be truer when it comes to news articles that describe scientific innovations such as the launch of high-resolution satellites and the imagery they collect. So then the focus of this month’s Geospatial Tip of the Month (GTM) is debunking a series of misleading articles about high-resolution satellite imagery that have been posted on ‘media’ websites. These articles, and the many like them, lead to incorrect understanding of and thus false hopes for consumers and companies looking for satellite imagery that may remedy their immediate problems.

Myth 1 – Satellite Imagery Can Solve Crimes

If you read through this article posted on The Week, you will find out about a UK based firm, Air & Space Evidence Ltd of London, which purports to use satellite imagery to solve all sorts of crimes. In this article, you will see this admission:

The University of London academics … emphasised that not all cases could be solved through satellite data, but said that the range of evidence available was increasing in quantity and sophistication.

And the article goes on to say this:

“We don’t have data for every minute yet, so solving a murder case or a hit and run is difficult,” Mr Purdy said. “But if you’re looking at something that took place over a certain day or week, we might be able to play a key role.”

Well, here is the reality of the situation when it comes to satellite imagery: it is highly unlikely that a satellite would collect imagery of a specific location on a specific day or even week. And here is why, a satellite only ‘sees’ each location on the planet on average every 2 to 3 days; but this only part of the story. While a satellite can see each location every 2 to 3 days, it can only image far less than 0.1% of the land mass it can see each day given the time it takes to image the ground and the onboard storage capacity. So if you were to look in the archive over a given location, let’s say my house in Boulder, CO, you would only see 13 images with 1-meter or better resolution collected in 2014, and of these, only 5 of them actually show my house as the rest are cloudy.

The long and the short is this, in some very, very limited cases, satellite imagery might help solve crimes. But in the vast majority of cases, we will not have data, which gives crime victims visions of false hopes. To give you a sense, I have sold satellite imagery for nearly 10 years now, and in those 10 years, I have only found imagery twice from a specific time period and location for a client.

Myth 2 – Satellite Imagery is Real-Time

In this article by Motherboard, the claim is made that with the new ~1-meter satellites launched by Skybox Imaging (now owned by Google), that there will be real-time satellite imagery. Here is what they claim:

So, when you get that panicked feeling mid-flight that you forgot to turn off your coffee maker before leaving on vacation, Google will have resolution adequate enough for you to see a recent image of your slowly singeing house. You could also hypothetically pull up Google Maps and see a real-time image of your actual plane, rather than a blue dot, soar further and further away from your fiery abode.

The article actually goes on to debunk itself as satellites cannot offer real-time imagery and here is why:

Google’s Skybox intends to launch a constellation of 24 satellites by 2018, which will survey the globe by taking pictures of its entirety three times a day. This too will eventually, undoubtedly, upgrade the picture quality on its map applications. Google’s own birds can capture 90-second video clips and imagery at 30 frames per second.

So then if you extrapolate from the quote above, assuming Google/Skybox even launches 24 satellites, each location on the planet could be recorded a maximum of 4.5 minutes per day. This would assume recording mode over all locations a satellite can see which would again exceed its onboard storage – but either way, 4.5 minutes out of 1,440 minutes in a day is far from real-time.

Then there is the issue that satellites have to beam data back to the planet and it needs to be processed, so that even this 4.5 minutes of possible video is certainly not real-time. In the foreseeable future at least, there will always be some lag between satellite collections and the ability to view this data.

Myth 3 – Satellite Imagery Can See Objects on Your Desk

In our third news piece published by the UK site, The Independent, the following claim is made by an interviewee:

These two images will illustrate the difference in resolution a satellite can offer versus Google Maps much better than anything I can write here. This first image is the maximum resolution a satellite can currently offer at 30-cm. Check out the people on the paths, patios and lawn – they are the small black specks – and then the cars to compare to the next image. A 30-cm color image collected by WorldView-3 over the Washington Monument, Washington D.C. on September 28, 2014. (Image Courtesy: DigitalGlobe)

This second image is the Washington Monument as seen in Google Earth at maximum resolution. You can see much more detail in this image versus the WorldView-3 data above. For example, people are larger specks in this Google Earth image as are the windshields of cars. So it’s obvious you cannot see objects on a desk in WorldView-3 imagery, nor can you see them in a more detailed Google Earth image.

“If you imagine a satellite sat above your office then the old resolution could probably make out your desk. The new imagery – where each pixel measures around 31cm – can now make out what’s on your desk,” explains Clive Evans, lead satellite imagery investigator with LGC Forensics.

Let’s dissect the comment above sentence by sentence. The first claim made is that you could probably make out your desk. Well, that part might be true as you can see objects as small as a car in some detail and then objects smaller than that in limited detail. So if a desk was out on a lawn (by would obviously be weird) with a collection of objects like a lawn chair, a garbage can and a beach towel, you would likely see each of these objects as tiny specks; but certainly not in enough detail to determine what each speck in the imagery was. We can call this first sentence partially true then.

But the second sentence could not be farther from the truth. In the past, the maximum resolution for satellite imagery by US government regulation was 50-cm. Now you can receive imagery with 31-cm resolution (at best). So while this will show more detail than 50-cm resolution, it will not show you objects on a desk unless they were large (perhaps a foot by a foot) and of a different color than the desk. And even if you did ‘see’ these objects, they would literally be a single pixel of color.

For example, let’s say you had a wooden desk that was 4-feet long by 2-feet wide. In a 31-cm satellite image, the desk would be made up of 8 beige colored pixels (or squares of color), not enough detail to tell if it was a desk or beige colored towel. Now if you placed a black laptop on its surface, that would be a single pixel of black against 7 pixels in a rectangular format; again, is that a laptop on the desk or a beige beach towel with a black pattern on it? It would be 100% impossible to determine this with any certainty in 31-cm data, myth debunked.

Wow, this has been so much fun that in next month’s GTM, I will feature 3 more debunked satellite imagery myths!

Do you have an idea for a future GTM? If so, let me know by email at brock@apollomapping.com.

Brock Adam McCarty

Map Wizard
(720) 470-7988
brock@apollomapping.com

This entry was posted in The Geospatial Times and tagged , , , Bookmark the permalink.

One Response to GTM – Debunking Satellite Imagery Myths, Part I

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.

    The Geospatial Times Archive