IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
Map-making from imagery, Amateur astronomy, Lunar map
JRehling
post May 21 2015, 06:42 PM
Post #1


Senior Member
****

Group: Members
Posts: 2530
Joined: 20-April 05
Member No.: 321



I'm simply going to share a personal anecdote whose main point will be of no surprise to those with similar experiences:

I've tasked myself with making a lunar map using imagery that I myself gathered with a small telescope and a digital camera. In principle, this can be done with one image of the full Moon, but my goal was to get imagery with the terminator located at all longitudes across the lunar face, so that the merit of shadows revealing topography could be enjoyed. It is easy to find maps of this kind, but I thought it would be fun to make my own. And it is fun, but it's also a hell of a lot of work.

I had the extraordinary luck of clear California skies 13+ nights in a row precisely when I decided to begin the project. I got full-disk images 11 nights in a row, and then "all" I had to do was process the results. The process is still going.

Projecting a global image into a cylindrical map is extraordinarily sensitive, at the margins, upon having the geometry just right. (The pixel representing a pole becomes a line as wide as the map, whereas the pixel right next to it becomes something much smaller.) But despite clear skies, I had much more than one pixel of distortion across my images due to imperfect seeing. (That's an issue that spacecraft don't have to worry about, but they have other issues, such as motion.) This made it challenging to register my global images with adequate precision.

My whole pipeline, now in progress, is like this:

1) Take several images of the Moon.
2) Montage them into a global view for that night. (This is already challenging when seeing is bad.)
3) Geometrically align the global image into a view my map-projecting software (which I wrote) expects.
4) Using data on that date-time's lunar libration geometry, project it into a cylindrical map.
5) Resize and reorient this to "pin" it to an existing cylindrical map from LRO. This compensates for some error.
6) Select strips near the terminator from the 11 pinned maps.
7) Adjust brightness across the strips to achieve consistent illumination throughout.

I'm on (5) now, and may hit more unpleasant surprises as I go. The best I can achieve will still distort the positions of some features, but I hope to get a pretty attractive and pretty accurate map with about 2.5 km/pixel resolution covering about ±70° from (0°, 0°).

I could potentially expand this into a more detailed map, but the burdens would only increase.

I also hope to generate a map of Mars during the 2016 season next spring. That will be a much coarser-resolution product, but also a lot less post-processing work.

Go to the top of the page
 
+Quote Post
Phil Stooke
post May 21 2015, 08:51 PM
Post #2


Solar System Cartographer
****

Group: Members
Posts: 10122
Joined: 5-April 05
From: Canada
Member No.: 227



I have done a lot of work like this over the years, and one thing I have found is that the poles are best dealt with separately. Typically, I would take all my images of the north polar region, and project them into a polar azimuthal projection out to about 40 degrees from the pole. When i am satisfied that it looks OK, I will reproject that into a cyindrical projection. Then, I will do the same for the south pole. At that stage i will have a cylindrical projection with the equatorial/midlatitude strip missing. Then all remaining images are projected to fill the gap.

This is the process I used to map Itokawa and Eros, both of which were documented on this site.

Phil


--------------------
... because the Solar System ain't gonna map itself.

Also to be found posting similar content on https://mastodon.social/@PhilStooke
NOTE: everything created by me which I post on UMSF is considered to be in the public domain (NOT CC, public domain)
Go to the top of the page
 
+Quote Post
JRehling
post May 21 2015, 11:13 PM
Post #3


Senior Member
****

Group: Members
Posts: 2530
Joined: 20-April 05
Member No.: 321



Good advice, Phil!

Something I hadn't been so conscious of until I began this exercise is how substantial lunar libration is, so during the half-month in question, the north pole was easily visible (near the limb, of course, but visible) and the entire south polar region hidden. So, the option of making a south pole map would require new imagery from another time. But neither pole will ever be exceptionally favorable, so weird, smudgy, streaky images is the best I'd ever get.

I have been intrigued by the idea of doing Phil-o-vision of the limb, after compensating for the overall global curvature. The topography along the limb can be quite intriguing.
Go to the top of the page
 
+Quote Post
Gerald
post May 21 2015, 11:22 PM
Post #4


Senior Member
****

Group: Members
Posts: 2346
Joined: 7-December 12
Member No.: 6780



- As an additional step you might need a correction for optics and/or camera distortion; determining the first radial parameter K1 of the Brown model should improve results.

- With quite some number crunching you may achieve subpixel registering. I've done so, e.g. for the New Horizons processing (with about 0.004 pixels accuracy to get near the geometric camera distortion); the by far most computer power consuming part.
This should help registering the exact location of the moon, as well as the rotational phase.
Minimizing RMS errors for a 30x30 pixel tile, with bilinear interpolation, if necessary, should do a good job for moon images, provided you're able to constrain relative rotation and position a priori sufficiently accurate (up to a few pixels, and up to a few degrees). Check all 1-pixel displacements and -rotations, then switch to a hill climbing algorithm for subpixel registering.
For images with significant respective differences in absolute brightness for the selected tile, first apply hipass filtering.
Go to the top of the page
 
+Quote Post
Bjorn Jonsson
post May 22 2015, 12:57 AM
Post #5


IMG to PNG GOD
****

Group: Moderator
Posts: 2250
Joined: 19-February 04
From: Near fire and ice
Member No.: 38



Very interesting project. One possibility to compensate for the atmospheric distortions near the limb might be to warp the images to the 'correct' shape before reprojecting them (if the correct shape can be determined to high accuracy).

An additional problem that can get severe very near the limb is topography if you assume that the Moon is spherical or an ellipsoid. This is probably not a problem for the Moon unless the images are of really high resolution. In contrast, for bodies like Mimas (and especially Vesta and more irregular bodies) it is usually necessary to reproject the images onto a 3D shapemodel of the target.
Go to the top of the page
 
+Quote Post
JRehling
post May 22 2015, 05:29 PM
Post #6


Senior Member
****

Group: Members
Posts: 2530
Joined: 20-April 05
Member No.: 321



I think one of my most significant flawed assumptions was that the distance from the Earth to the Moon is essentially infinity. It's not, and so the limb isn't exactly at 180° from the sub-Earth point. This is most evident when a very thin crescent is observed, and the limb is illuminated somewhat less than 180° around.

The problem introduced by imperfect seeing is that, even though I shot video, no frame has an undistorted image of the whole Moon. Horizontal shifts in position occur continuously, like waves rippling across a surface. To get even one good picture of the Moon on a night like that, I'd need to perform elaborate composition of portions of different images. On ideal nights, this doesn't happen to a visible extent, but it probably take me years to gather all the imagery on perfect nights!

For a smaller object like Saturn, occasional frames turn out perfect even on a night with poor seeing, because the whole planet ends up being shifted in about the same way at the same time. (Although the vast majority do not.) But the Moon is far too big for that.

Gerald, that's an outstanding technique! I could probably benefit from that.
Go to the top of the page
 
+Quote Post
Bjorn Jonsson
post May 22 2015, 07:16 PM
Post #7


IMG to PNG GOD
****

Group: Moderator
Posts: 2250
Joined: 19-February 04
From: Near fire and ice
Member No.: 38



QUOTE (JRehling @ May 22 2015, 05:29 PM) *
I think one of my most significant flawed assumptions was that the distance from the Earth to the Moon is essentially infinity. It's not, and so the limb isn't exactly at 180° from the sub-Earth point.


Yes, this is significant near the limb. I know this from my experience when reprojecting images of the Galilean satellites where I sometimes had to use global images obtained at a range of ~300,000 km and farther out to fill small gaps in my maps of these bodies.

Go to the top of the page
 
+Quote Post
scalbers
post May 22 2015, 07:35 PM
Post #8


Senior Member
****

Group: Members
Posts: 1620
Joined: 5-March 05
From: Boulder, CO
Member No.: 184



I think with the scale of images you'd likely be working with that the error introduced in lunar images wouldn't be too bad. However it is possible to correct for this with the right equations added to the remapping procedure. I had used a reference called "Map Projections: A Working Manual" (1987) to learn about this.

When the angular diameter of the target increases to around 3 degrees this becomes more noticeable in cases I've worked with.


--------------------
Steve [ my home page and planetary maps page ]
Go to the top of the page
 
+Quote Post
JRehling
post May 23 2015, 05:55 PM
Post #9


Senior Member
****

Group: Members
Posts: 2530
Joined: 20-April 05
Member No.: 321



Thanks for sharing your wisdom, Bjorn and Steve, in addition to those who commented before!

I'm en route to producing a passable map, but I can already see how a better version can result if I go back to scratch.

It's remarkable how much could go into making a good lunar map from telescopic images if one really pursued the project to its full extent. It would take mosaicking hundreds of images, for starters, whereas I'm trying to get away with only 35 or so (about 3.5 per night for 10 or 11 nights). And then very careful sampling of the frames to compensate for imperfect seeing. And then all of the other wisdom that people have shared.
Go to the top of the page
 
+Quote Post
JohnVV
post May 24 2015, 01:39 AM
Post #10


Member
***

Group: Members
Posts: 890
Joined: 18-November 08
Member No.: 4489



without using something like ISIS3

Hugin works very well for mosaicing

you can create tie points manually or have the software find them


as above for the poles
do 90N-60N and 60S to 90S in polar stereographic
and 60N to 60S in Mercator

then combine them
Go to the top of the page
 
+Quote Post
JRehling
post Sep 9 2015, 12:24 AM
Post #11


Senior Member
****

Group: Members
Posts: 2530
Joined: 20-April 05
Member No.: 321



After much tinkering, I've realized a passable (to me) version of a map of the lunar nearside, made with my own imagery.

As mentioned earlier, the challenge was to composite many images so that locally, areas in morning shadow were used almost everywhere, making a much nicer product than one gets from projecting a full Moon onto a plane.

The challenges were many that are familiar to people who process space images, with some added bother due to atmospheric conditions. And, incidentally, a telescope so old that you couldn't even import things from China when it was made.

Here is a version jpg'd to fit on the board, with gridlines at 30° increments. Another version with major features labeled is in the works. Thanks to everyone for the tips offered earlier this year.
Attached thumbnail(s)
Attached Image
 
Go to the top of the page
 
+Quote Post
hendric
post Sep 9 2015, 04:01 PM
Post #12


Director of Galilean Photography
***

Group: Members
Posts: 896
Joined: 15-July 04
From: Austin, TX
Member No.: 93



What is your optical setup? Looks like a very nice map. Are you using any of the video stacking programs to automatically choose the best frames/subframes, or doing it all manually (!)?


--------------------
Space Enthusiast Richard Hendricks
--
"The engineers, as usual, made a tremendous fuss. Again as usual, they did the job in half the time they had dismissed as being absolutely impossible." --Rescue Party, Arthur C Clarke
Mother Nature is the final inspector of all quality.
Go to the top of the page
 
+Quote Post
JRehling
post Sep 9 2015, 05:45 PM
Post #13


Senior Member
****

Group: Members
Posts: 2530
Joined: 20-April 05
Member No.: 321



The telescope was a 3" Jason telescope with 960mm focal length almost 40 years old, with no tracking and considerably degraded optics! The camera was an Orion Starshoot Solar System Color camera.

A great impediment, I found, was the way atmospheric activity (on many nights, but not all) made the image waver like a flag blowing in the wind. Or, on bad nights, go in and out of focus, locally. I lined up and merged subframes manually, suspecting that the nearly ubiquitous spatial distortions would trick any algorithm into misbehaving. Basically no two frames ever showed the same appearance; there was always something wavering.

I never noticed this while observing with my eye because one only sees what the fovea is aimed at at any given moment and thus, one misses that broad frequency (~10 minutes of arc) wavering. Smaller targets (basically, Jupiter or anything smaller) are below that frequency and seem to move left/right/up/down as a whole and not show so much distortion within the object's image.

This could be addressed by taking the best observations for each phase over, say, an entire year or more, but I took all of this imagery on just 11 consecutive nights.

The relatively limited resolution was a mixed curse/blessing, because it mitigated the mosaicking work.

I'll be writing the whole process up in detail in a blog post to appear soon.
Go to the top of the page
 
+Quote Post

Reply to this topicStart new topic

 



RSS Lo-Fi Version Time is now: 19th March 2024 - 04:45 AM
RULES AND GUIDELINES
Please read the Forum Rules and Guidelines before posting.

IMAGE COPYRIGHT
Images posted on UnmannedSpaceflight.com may be copyrighted. Do not reproduce without permission. Read here for further information on space images and copyright.

OPINIONS AND MODERATION
Opinions expressed on UnmannedSpaceflight.com are those of the individual posters and do not necessarily reflect the opinions of UnmannedSpaceflight.com or The Planetary Society. The all-volunteer UnmannedSpaceflight.com moderation team is wholly independent of The Planetary Society. The Planetary Society has no influence over decisions made by the UnmannedSpaceflight.com moderators.
SUPPORT THE FORUM
Unmannedspaceflight.com is funded by the Planetary Society. Please consider supporting our work and many other projects by donating to the Society or becoming a member.