Voyager mosaics and images of Jupiter, A fresh look at some ancient stuff Options

[4th rock from th...]

The trouble with this kind of images is overexposure. You don't have the full dynamic range on any channel, so you can't balance them visually.

You can try to calibrate colors mathematically, but that is hard to do with vidicon based imagery.

Another possible option would be to use the rings as a "white balance target". Don't know if they are 100% white, but the principle is the same. You can use that part of the image to balance the original channel data. Something like doing image histogram equalization using that part as reference....
At least for the midrange tones, you should get relatively accurate results. Good enough to distinguish "more red" or "more blue" areas, certainly.

[geckzilla]

Another possible option would be to use the rings as a "white balance target".

That's exactly what I did. I should have been more careful because the whole thing is slightly too blue. Oops. Went ahead and fixed that. I guess there's not much more that can be done.

[Phil Stooke]

Saturn's icy rings are white - pretty close to it anyway - but Jupiter's are made of dust knocked off small rocky inner moons, possibly polluted by sulfur and other material from Io's volcanoes. It would be dangerous to assume they are white. I have not looked up data on the rings but it must be out there.

Phil

[4th rock from th...]

Looking at the image yes, both channels look linear over the rings. Perhaps there are some registration errors, at least I see some color banding in parts of the ring's edge.
You can run a median filter to discard any color variation smaller than 2 or 3 pixels.

The key here is to know the backlit color of the rings :-) and readjust the image for that :-)

[geckzilla]

Is it typical to shift parts around to fit them? The rings are slightly off, but that is because I felt the limb was more important to have registered more closely because the rings are already kind of monotone (whatever color they are, curious question, that) and diffuse. I could have them both registered if I cut the rings out and shifted them a pixel or two. I don't know what kind of actions are looked down upon for scientific images like these. At some point it goes from truth to truthiness.

[4th rock from th...]

You get mis-registration because spacecraft movement creates a parallax effect on sequential frames
You can also have geometrical distortions on the images (they come from vidicon cameras, not CCD imagers) that create apparent shifts, if the object moved across the field of view.

It's OK to shift parts manually in my opinion, but you must state that.
State what you corrected on each part, so that others can follow your steps and verify that you just didn't made things up

[Bjorn Jonsson]

Geckzilla - welcome to the forum. This is interesting work and better than e.g. this welll known image where the color channels aren't properly aligned.

I agree that it's OK to shift parts around a bit (or even 'warp' individual images a bit) to get things to fit as long as you clearly state that you did this. For satellite transits it's also common to process/shift the satellite images separately from the background planet when doing color composites - otherwise it would usually be impossible to get color images where both the satellite and the planet look OK.

The raw Voyager images are also a bit distorted and this can be a significant effect in some cases, especially when mosaicking but also when aligning color channels.

[meryon]

Hello all !

This thread is totally insane ! Very beautiful photos.

I have a question about the color treatment.

I have downloaded the TAR file from the Voyage database. I presume that you work with IMG files.

But how do you extract the three colors from this file? When I open it in NASAView it is only in black and white and I don't find an operation mode to extract the three channels.

Thanks for all an specially for your amazing images !

Ben

[elakdawalla]

Meryon, virtually all deep-space cameras have monochrome detectors. They produce grayscale images. For color imaging, they must take sequential photos through three different-color filters: red, green, and blue for approximately true color. You may wish to visit my pages on space imaging and image processing tutorials to learn more, and here's a blog entry where I talk about some of the problems of making color pictures from three grayscale ones taken from a moving spacecraft.

[JohnVV]

to add to emiy 's post
NASAView outputs 8 bit indexed gif images
basically useless for converting the three red, green,blue images into a color image

i myself use ISIS3 but it dose NOT run on Microsoft Windows
http://isis.astrogeology.usgs.gov/index.html

you NEED RHEL6.6 or one of the free versions of that Linux operating system
( i use ScientificLinux 6.6 )

BUT Bjorn Jonsson has a cool tool that DOSE run on windows called "img2png"
but you NEED to use the Microsoft windows terminal "cmd.exe"
http://www.unmannedspaceflight.com/index.php?showtopic=4979

converting a "infrared" and "ultraviolet" and "green" images into "RED , GREEN , BLUE " image is a bit complex
besides the links above
http://www.unmannedspaceflight.com/index.php?showtopic=7671



some people use Matlab *.m files ( or the free GPL'ed Octave program) to recalculate the pixel values from what the camera sees to what the human eye sees
Matlab ( or Octave) is a GREAT mathematical program .

[meryon]

OK! Thank you so much both of you.

I will read attentively these pages.

I'm graphist in VFX industry and my main program is Photoshop. Do you know if is there any plugin to process IMG file ?

See you soon

[JohnVV]

10 years ago there was a photoshop plugin
if it is still around ?????

the voyager images are normally in a IMQ ( notice the Q is not a G )
a compressed image format

but not all of archives are

voyager1 &2 reworked cd's
http://pds-imaging.jpl.nasa.gov/data/voyag...SATURN/C3465XXX
is in img
but
http://pds-imaging.jpl.nasa.gov/data/voyag...saturn/c3465xxx
is in IMQ


BUT the good news is for MS windows GDAL will open and convert the img voyager file BUT you MUST !!! have the MATCHING pds lable the LBL file
http://www.gdal.org/



for example this image of saturn
http://pds-imaging.jpl.nasa.gov/data/voyag...3465138_RAW.IMG
http://pds-imaging.jpl.nasa.gov/data/voyag...3465138_RAW.LBL

once the Microsoft version of GDAL is installed and it works
You can do this in "cmd.exe"
-- for DEFAULT settings you can leave out some commands

CODE

gdal_translate C3465138_RAW.LBL C3465138.tiff

both the img and lbl have to be in the same folder
you then get this 8 bit image


but you still have the Reseaus marks ( the black dots )
and some lens distortion

[Phil Stooke]

The LBL file will give image dimensions somewhere - it can take some searching. It might say 1024 lines, 1024 samples and 8 bits/pixel, for instance. So you can use Photoshop to open that file as a RAW image, entering those dimensions and using the guess function to guess the header size if necessary. I open IMG files all the time in Photoshop just as raw images.

Phil

[JohnVV]

the issue with the voyager images is that on EACH line of the image there is a 224 byte prefix

to quote the lbl

DESCRIPTION = "This image is the result of decompressing
the corresponding source image C3465138.IMQ as found on one of the Voyager ISS
CD-ROMs. It was expanded using VICAR routine VGRCDCOPY on a Compaq Alpha
running OpenVMS. It contains the raw image as an array of 800x800 unsigned
bytes, with a prefix of 224 bytes on each line.

See file DOCUMENT/PROCESSING.TXT for more information about the image
processing."

opining it as a raw in( photoshop,imagemagic,gmic,...) is not really a option

for each of the 800 lines you need to skip 224 bytes ( for each line)

as i recall the real to real tapes used a video image format
this was the 70's and 80's
the age of the TSR80 and commodore 64 and 8 inch floppy's

[djellison]

i myself use ISIS3 but it dose NOT run on Microsoft Windows
http://isis.astrogeology.usgs.gov/index.html

you NEED RHEL6.6 or one of the free versions of that Linux operating system

No you don't. ISIS3 is also available for OSX.


meryon - if you're on a PC - you need to go IMG2PNG -> Photoshop. If you're on a Mac - then an ISIS3 pipeline is probably your best bet.