Jump to content

Stargazers Lounge Uses Cookies

Like most websites, SGL uses cookies in order to deliver a secure, personalised service, to provide social media functions and to analyse our traffic. Continued use of SGL indicates your acceptance of our cookie policy.

Sedna

Using machine learning to "enhance" the hemisphere of Pluto not yet mapped in high resolution (with your help!)

Recommended Posts

When New Horizons flew by Pluto and Charon in 2015, each body had one hemisphere that was never mapped at higher resolution than what we see in the images below. Because Pluto and Charon rotate relatively slowly (6.4 days), only one hemisphere of each was facing New Horizons during its closest approach.

image.thumb.png.121988c787d5096c61dc271f4c47aec4.png

 

image.png.1c0df6e745e2919541303557372c4174.png

Will we ever see the other face of Pluto (or Charon) mapped in high resolution? The face of Pluto that was mapped in low-resolution is the hemisphere that is tidally locked with (permanently facing) Charon. From New Horizons low resolution imagery, we see that it is very different from the hemisphere with the "heart" (giant glacier) that was mapped in high resolution (see below). Unlike that hemisphere, the Charon-facing hemisphere features strange "brass knuckle" spots and mysterious dark swirls ...

PressImageContext_12-10-15-624x365.jpg

It's a shame we might never see this hemisphere in high resolution. However, for artistic purposes, machine learning might offer a solution. You might already know that machine learning can be used to "colorize" black and white photos by learning which colors go with which patterns (you can try this with any black and white here and read more about it here). Similarly, machine learning can take crude drawings of faces and turn them into high resolution "paintings" by learning which high spatial frequency textures and colors are often paired with which low spatial frequency patterns in the drawing (you can read about this here).

image.png.403f6ab910ac5a8310e70c982a0d2da0.png

Based on the same principle, a machine learning algorithm could likely learn which high spatial frequencies are statistically more likely to pair with low spatial frequencies in high resolution images of Pluto from New Horizons. After being trained on these images from Pluto's high res hemisphere, the machine learning algorithm could then do some guess work and apply high frequency textures to the low resolution images (currently containing only low frequency textures) of the Charon-facing hemisphere. The result should give us some beautiful guesswork: images of hitherto unseen craters, canyons, and ice dunes on the "far side" of Pluto. Of course, these images would be only that--a guess. But they might be convincing enough to quench our thirst for what New Horizons could never provide while exceeding the quality of the best space art.

At this point, you might be hoping to see what a neural network could dream up for Pluto. Sadly, I lack the technical proficiency to implement this myself. So I'm hoping that some of you could help me (and everyone who'd like to see this) implement this project using machine learning. All you programmers, image processors, and data scientists out there--am I way off or is this doable? If the latter, would someone like to stand up and seize the glory?

If you're interested in this project, please reply here or private message me. Thank you!

  • Like 2

Share this post


Link to post
Share on other sites

As an art project, it is certainly a fun idea. As  the Charon-facing hemisphere is different from the other hemisphere, statistics picked up by deep learning may not provide any reliable guesses of what the link is between low and high frequency detail on the Charon-facing side. It might be an interesting project for students in the upcoming Computer Vision course I teach.

  • Like 1

Share this post


Link to post
Share on other sites
9 minutes ago, michael.h.f.wilkinson said:

 It might be an interesting project for students in the upcoming Computer Vision course I teach.

Ah, that sounds like a great idea! If you assign it to them, would you post the best projects here (with their permission)?

Share this post


Link to post
Share on other sites
12 minutes ago, Sedna said:

Ah, that sounds like a great idea! If you assign it to them, would you post the best projects here (with their permission)?

Sure, they would be proud, I would think. Do you have a particular data set in mind to work on?

 

Share this post


Link to post
Share on other sites
6 hours ago, michael.h.f.wilkinson said:

Sure, they would be proud, I would think. Do you have a particular data set in mind to work on?

 

No date--whenever it happens, I'll be thrilled :D

Share this post


Link to post
Share on other sites
On 30/01/2019 at 17:41, Sedna said:

Sure--the raw data is all here and images already processed by NASA are here. @michael.h.f.wilkinson, if you need anymore help, happy to assist!

Not sure raw data in JPEG format will help too much, but I could let students search within the NASA PDS. Not entirely sure whether this will work, but I will see

 

Share this post


Link to post
Share on other sites
4 hours ago, michael.h.f.wilkinson said:

Not sure raw data in JPEG format will help too much, but I could let students search within the NASA PDS. Not entirely sure whether this will work, but I will see

 

Ah, sorry about that, didn't realize they were JPEGs. Okay, how about trying here instead. If you scroll down to Pluto Encounter, everything under LORRI (New Horizon's high focal length black and white camera) and MVIC (the low focal length color camera) should be relevant. Please let me know if you need any more assistance!

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...

Important Information

By using this site, you agree to our Terms of Use.