IPB

Welcome Guest ( Log In | Register )

 
Reply to this topicStart new topic
Voyager Images and Isis3, applications and methods
JohnVV
post Jun 5 2016, 07:41 PM
Post #1


Member
***

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



read through and run the examples in the Workshop
https://isis.astrogeology.usgs.gov/IsisWork...hp/IsisWorkshop

isis3 can be a bit daunting .
First the linux requirement ( there is a mac port)
the rhel 6 build runs just fine on ( i have tested )
CentOS 6.8
ScientificLinux 6.7 ( will be 6.8 soon )
Opensuse 13.2 and 42.1
-- WARNING on Fedora - has issues

Then the old Voyager data

Those that have been fallowing any of the reprocessing threads have seen the images
missing lines and reseaus marks and it is 8 bit

and searching for data can be a problem
unless you are just garbing it all

for this post i am using a few Blue,Green and Orange ( no red filter) ( orange and blue make red )
and using the pds atlas
http://pds-imaging.jpl.nasa.gov/search/
-- WARNING!!!!!!
this archive sends isis3 incompatible "img" files !!!
the wget script needs to be "CORRECTED!!!


for example the "atlas_wget_script" give this
CODE
wget -O c1113006.img 'http://pds-imaging.jpl.nasa.gov/pds/prod?q=OFSN+%3D+/data/voyager//vg_0010/extras/pds/neptune/c1113xxx//c1113006.img+AND+RT+%3D+RAW'
wget -O c1113017.img 'http://pds-imaging.jpl.nasa.gov/pds/prod?q=OFSN+%3D+/data/voyager//vg_0010/extras/pds/neptune/c1113xxx//c1113017.img+AND+RT+%3D+RAW'
wget -O c1113039.img 'http://pds-imaging.jpl.nasa.gov/pds/prod?q=OFSN+%3D+/data/voyager//vg_0010/extras/pds/neptune/c1113xxx//c1113039.img+AND+RT+%3D+RAW'

these img files can not be opened with "voy2isis"
You need the imq ( Q on the end )
a simple edit of the script will solve this

--fixed--
CODE
wget -O c1113006.imq 'http://pdsimage.wr.usgs.gov/archive/vg1_vg2-n-iss-2-edr-v1.0/vg_0010/neptune/c1113xxx/c1113006.imq'
wget -O c1113017.imq 'http://pdsimage.wr.usgs.gov/archive/vg1_vg2-n-iss-2-edr-v1.0/vg_0010/neptune/c1113xxx/c1113017.imq'
wget -O c1113039.imq 'http://pdsimage.wr.usgs.gov/archive/vg1_vg2-n-iss-2-edr-v1.0/vg_0010/neptune/c1113xxx/c1113039.imq'


the voyager " *.imq " files are gziped compressed archives

isis3 has BUILT IN batch support -- very handy

to ingest into isis3 and also get a "preview" image i use this script
CODE
#!/bin/sh
ls *.imq | sed s/.imq// > cube.lis
voy2isis from=\$1.imq to=\$1.cub -batchlist=cube.lis
spiceinit from=\$1.cub -batchlist=cube.lis
voycal from=\$1.cub to=\$1.cal.cub -batchlist=cube.lis
isis2std from=\$1.cal.cub to=\$1.png minpercent=0.1 maxpercent=99.9 -batchlist=cube.lis


i have now a calibrated and geo-referenced image
these images


and they need work ( the pointing data is OFF - hay this is from Voyager after all )


( preview of "c1113006" using isis3's "Grid " program )


i will worry about that later

the spiceinit'ed and voycal images are now 2.5 Meg 32bit float images -- "c1113006.cal.cub"
have a look in Qview
CODE
qview c1113006.cal.cub



-- if you are going to use the built in isis3 tools then leave the NULL pixels and do not run stretch

-- for using gimp or other tools

the missing lines are "NULL" data ( px value -3.8 E^38 )
and some low instrument values that are negative pixel values
-- for cleaning up and exporting the images to Gimp2.9.3 ( or Krita )these need to be zero

a handy isis3 tool is "stretch "
CODE
stretch from=c1113006.cal.cub to=c1113006.cal1.cub null=0.0 lis=0.0


on to exporting it to a RAW image so it can be cleaned up and then moved back into isis3
best done in a new sub folder
CODE
cubeatt from=c1113006.cal1.cub to=c1113006.cal1.raw+BSQ+detached

you now have a raw and detached headers
( rename c1113006.cal1.raw.cub to c1113006.cal1.raw )
Go to the top of the page
 
+Quote Post
JohnVV
post Jun 5 2016, 08:16 PM
Post #2


Member
***

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



on to the next bit

this will require some other linux tools ( also build on windows )

Gmic
http://gmic.sourceforge.net/

DEVELOPMENT!!! version of Gimp 2.9.3

handles 16 bit and 32 bit image depth
-- this really is not hard to build and NOT mess up your centos or opensuse install
-- just a bit complex
https://github.com/GNOME/gimp
and using this guide
http://wiki.gimp.org/wiki/Hacking:Building/Linux
and that guide is based in this
http://lightningismyname.blogspot.com/p/compiling-gimp.html
and using the "config.site" for autotools
-- a script that makes that file is here
https://drive.google.com/file/d/0B6ZYAd08tZ...iew?usp=sharing
-- works fine

build gmic using gimp2.9.3
( or use the terminal version of gmic)

gmic is being used to fill in the lines ( there is also a lowpass way inside isis3 ) but i like this better

and the gimp "heal" tool is being used for the "reseaus marks"
( there is a built in tool in isis3 but it dose not give good results )

----------or Krita -----------------
or install Krita this editor already has Gmic-gimp plugin built into the program and is in most repos !!!
for opensuse "calligra-krita"
https://krita.org/

you will still need the TERMINAL version of Gmic, but the prebuilt WILL work , just NO display like Imagemagick has

-----------------------------

-----------------
after renaming the *.raw.cub to *.raw
open the new lbl file in a text editor ( gedit or kate) "c1113006.cal1.raw.lbl"

at the beginning is this
QUOTE
Group = Dimensions
Samples = 800
Lines = 800
Bands = 1
End_Group

Group = Pixels
Type = Real
ByteOrder = Lsb
Base = 0.0
Multiplier = 1.0
End_Group
End_Object


so it is a "real" -- 32 bit float image that is 800x800 px

convert the raw to a tiff using Gmic
CODE
gmic c1113006.cal1.raw,800,800 -o c1113006.cal1.F.tiff

tiff's can be many formats for a floating point i add a "F" to the name

screenshot of "c1113006.cal1.F.tiff" in gimp2.9.3 DEV


now we have the 32 bit image in gimp
the next thing to do is remove the horizontal lines of missing data

the NULL values were set to "0" using "stretch" a bit back
so the lines have a value of 0

filters / G'Mic
then in the dialog open "repair" and click on the "inpaint ( morphological )
and change the red masking color to black "0"
and click "ok"


and the lines are gone

onto the "reseaus marks"
use the heal tool" crossed "band-aids" and set the brush Opacity (transparency) to 15 %
and use the hardness 25% brush
and a 16 pixel brush

and save
Go to the top of the page
 
+Quote Post
JohnVV
post Jun 6 2016, 12:25 AM
Post #3


Member
***

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



now all this can also be done using the tools in isis3
-- just my preference i like gmic and gimp better

on the cal that was NOT streched to remove null px

run "findrx" to locate the reseaus marks
https://isis.astrogeology.usgs.gov/Applicat...drx/findrx.html
-- corrects the listing in the header

then run " remrx " to remove then
https://isis.astrogeology.usgs.gov/Applicat...emrx/remrx.html

set it to fill then with NULL and a box size of about 5x5 or 7x7


then fill in NULL pixels using a lowpass

-- WARNING javascript and needs to run a app in the sandbox
https://isis.astrogeology.usgs.gov/IsisWork...-in_NULL_pixels

from the level 1 workshop
https://isis.astrogeology.usgs.gov/IsisWork....php/Level1_ISS

run "lowpass"
https://isis.astrogeology.usgs.gov/Applicat...ss/lowpass.html
and there is a single check box that needs setting "filter pixels outside low and high "
CODE
lowpass filter=outside

and the boxcar size of about 11x11 or 13x13

-- side by side 8 bit copies
using gimp and using isis tools


gimp on the left isis3 tools on the right

the image on the right has some visible horizontal artifacts from the lowpass
gmic's fill looks more natural



next remapping and sub pixel registration
a solid surface WORKS better than CLOUDS
Go to the top of the page
 
+Quote Post
Astroboy
post Jun 6 2016, 06:17 PM
Post #4


Junior Member
**

Group: Members
Posts: 45
Joined: 27-August 14
From: Private island on Titan
Member No.: 7250



Having great success with the Isis directions... keep it up! Weird to finally be able to do this. As you said, the Isis lowpass filter is sort of not perfect, but converting the images to .png and adding a deinterlace filter fixed it, at the cost of some slight blurriness (which I honestly doubt I would notice unless I was told it was added).


--------------------
aka the Vidiconvict
Go to the top of the page
 
+Quote Post
JohnVV
post Jun 6 2016, 11:44 PM
Post #5


Member
***

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



on to aligning and remapping and registering

i said i was going to get back to this


we have two options that are easy
the image is open in gimp ( or still in a cub file)
this bit is trial and error so expect a few tries with moving it one or two pixels

isis3 has "translate" ( not language ) but move the image X number of samples or X number of lines
https://isis.astrogeology.usgs.gov/Applicat.../translate.html

measure the above 006 image in qview there is a measuring tool
it is aprox -136 samples and +40 lines off from the grid

run translate
then rerun on a test the grid program

-- you do not want the grid in the image
-- this is just to see if it is close

for gimp 2.9.3
you can run the grid program with the grid set to a value of 1 and background to 0
-- grid only -- no planet
CODE
grid from=c1113006.cal.cub to=grid.cub linevalue=dn dnvalue=1 bkgndvalue=dn bkgnddnvalue=0

then convert it to a tiff
CODE
gdal_translate grid.cub grid.tiff


open the grid and the image
copy/past the image into gimp
select the pasted layer as a new layer and set to lighten only


drag and drop the top layer until it is lined up
then delete the grid layer and flatten the image


and do this for all the image

now hopefully ( cross your fingers ) these are going to be "close "
a simple test to see just HOW far off they are in to remap( if needed) and join into a "TEST" image
-- DO NOT expect this to look great

-- for Gimp 2.9.3
time to ingest back into isis3
convert the last tiff to a raw image
CODE
gmic c1113006.F2.tiff -o c1113006.F2.raw

edit the lbl file to point to this NEW raw image and rerun "cubeatt "
CODE
cubeatt from=c1113006.raw.lbl to=c1113006.lev1.cub



isis3 has a tool to blend cubes into multi layer cubes "cubeit"
https://isis.astrogeology.usgs.gov/Applicat...eit/cubeit.html

needs a list
CODE
ls *.cub > list.lis

this is a orange ,green,blue blend with just the very BASIC of alignment ( and a TEST)
open in qview and select rgb and layer 1 ,2,3 and max the stretch



not too bad, there is some rotational offsets

that need fixing

this is where re-projecting and remapping comes in
cam2map ,map2map ,and map2cam and the use of control networks
Clouds do not work to well with the auto tools
https://isis.astrogeology.usgs.gov/Applicat...ontrol_Networks

( i am NOT going to get into "Radiometric_and_Photometric_Correction" yet )
https://isis.astrogeology.usgs.gov/Applicat...tric_Correction
Go to the top of the page
 
+Quote Post
JohnVV
post Jun 7 2016, 03:33 AM
Post #6


Member
***

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



i like to run tests before creating a control net ( they are time consuming )
and for small images like this might not be needed as long as the "rotational" data in spice is not off
and gas giants are a pain in the rear end for sub-pixel registering

so 3 images a orange( not converter to a red yet) green and blue camera image format

these need to become Simple Cylindrical for this
take a look in $ISIS3DATA/data/base/templates/maps/
copy and edit "simplecylindrical.map"
this is the one i am using
CODE
Group = Mapping
  TargetName         = Neptune
  ProjectionName     = SimpleCylindrical
  EquatorialRadius   = 24764000.0 <meters>
  PolarRadius        = 24341000.0 <meters>
  LatitudeType       = Planetocentric
  LongitudeDirection = PositiveEast
  LongitudeDomain    = 180
  Scale              = 5 <pixels/degree>
  MinimumLatitude    = -90.0
  MaximumLatitude    = 90.0
  MinimumLongitude   = -180.0
  MaximumLongitude   = 180.0
  CenterLongitude    = 0.0
End_Group

save it ( i rename it to simplecylindrical1.map ) in the folder with the 3 images
run "cam2map" and check "match map"
https://isis.astrogeology.usgs.gov/Applicat...ap/cam2map.html
CODE
cam2map from=c1113017.cub map=simplecylindrical1.map to=c1113017.map.cub matchmap=yes

then run "cubit" on all 3 new maped files
orange first then green then blue in the file list

from examining the 3 images the Orange is in the middle with green and blue on the left and right

lines up fairly well

on to "map2cam"
https://isis.astrogeology.usgs.gov/Applicat...am/map2cam.html
( you could also use "cam2cam and skip the simple cylindrical map )

CODE
map2cam from=ogb.cub match=c1113017.cub to=OGB.cam.cub

using the ORANGE image as the match image

side by side

on the left is the new remapped image and on the right is the post above test image

still needs some work BUT the color separation on the edges ARE A FACT OF LIFE
unless you use many images and make a full planet map

but some work can be done to lessen the color changes on the edges and make a RED layer ( instead of a orange )
Go to the top of the page
 
+Quote Post
Bjorn Jonsson
post Jun 7 2016, 11:12 PM
Post #7


IMG to PNG GOD
****

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



Big thanks for posting this, lots of very useful information here.

Have you experimented with the calibrated Voyager images (available from the PDS rings node) in ISIS3? If I understand correctly you are using the original (not calibrated) 800x800 pixel images in the examples above. The calibrated files are 1000x1000 pixels and the FOV is also slightly different, however the Voyager kernels that can be downloaded from the USGS ISIS website seem to assume that the FOV is as in the non-calibrated files.
Go to the top of the page
 
+Quote Post
JohnVV
post Jun 8 2016, 01:35 AM
Post #8


Member
***

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



I have basically stuck with the edr's with voyager
this data predates isis and parts of vicar
-- vicar is now on Github
https://github.com/nasa/VICAR

but this got me thinking and have a look
CODE
http://pds-imaging.jpl.nasa.gov/data/voyager/vg_0026/saturn/c3353xxx/c3353709.imq

opus -- ring node

http://pds-rings-tools.seti.org/opus/#/planet=Saturn&typeid=Image&missionid=Voyager&timesec1=1980-09-27T02:16&timesec2=1980-09-28T02:17&qtype-volumeid=contains&view=detail&browse=gallery&colls_browse=gallery&page=1&gallery_data_viewer=true&limit=100&order=time1&cols=ringobsid,planet,target,phase1,phase2,time1,time2&widgets=timesec1&widgets2=&detail=S_IMG_VG1_ISS_3353709_N

http://pds-rings.seti.org/volumes/VGISS_6xxx/VGISS_6105/DATA/C33537XX/C3353709_CALIB.IMG
http://pds-rings.seti.org/volumes/VGISS_6xxx/VGISS_6105/DATA/C33537XX/C3353709_CALIB.LBL

the calib is 800x800
BUT
the C3353709_GEOMED.IMG is 1000x1000

this is adding a lens warp to unbend the edges and corners
as you can see on this close up of Neptune c1133403.cub

you can see the frame edge mark for alignment in the bottom right corner

unfortunately the img's in the ring node error with
CODE
"No value or default value to translate for translation group
               [MissionName] in file
               [/mnt/USB.Drive/USB.P2/ISIS3/data/base/translations/MissionName-
               2DataDir.trn]"

i will have to manually edit and enter data into the labels
missing the mission name so i can not run spiceinit

However the edr imq files on the ring node are the same as the ones on the imaging node
http://pds-rings.seti.org/volumes/VG_0xxx/...ATURN/C3353XXX/

from the imq in the imaging node i can run spiceinit on it and georeferance it and remap






some fairly cool things can be done once images are Geo-referanced
this i did a bit back
Saturn's northpole ( Clear filter only )

see the hexagon !
Go to the top of the page
 
+Quote Post
Brian Burns
post Sep 3 2016, 09:09 AM
Post #9


Junior Member
**

Group: Members
Posts: 54
Joined: 7-July 16
From: Austin, Texas
Member No.: 7991



Thanks for the instructions. I decided to bite the bullet and try ISIS - I've got my virtual machine all set up, everything is installed, but I'm running into the same problem - the IMG LBL file missing the MissionName.

CODE
# ingest the pds file (combines LBL and IMG files into a cube file)
$ pds2isis from=C1625003_CALIB.LBL to=foo.cub

# view it - works
$ qview foo.cub

# append SPICE data to cube file
$ spiceinit from=foo.cub
**PROGRAMMER ERROR** No value or default value to translate for translation group [MissionName] in file [/home/bburns/Isis/isis/../data/base/translations/MissionName2DataDir.trn].

# check the embedded label - missing the MissionName
$ catlab from=foo.cub
Object = IsisCube
  Group = Archive
    DataSetId          = VG1/VG2-J-ISS-2/3/4/6-PROCESSED-V1.0
    ProductId          = C1625003_CALIB.IMG
    InstrumentHostName = "VOYAGER 1"
    InstrumentHostId   = VG1
    InstrumentName     = "IMAGING SCIENCE SUBSYSTEM - NARROW ANGLE"
    InstrumentId       = ISSN
    TargetName         = JUPITER
    MissionPhaseName   = "JUPITER ENCOUNTER"
  End_Group
End_Object
End


There's a Voyager-specific routine though -

voy2isis
Import PDS formatted Voyager EDR into Isis format cube
This program will take in raw image files from the Voyager spacecraft mission and import them into Isis formatted cubes. The input files are compressed Planetary Data System (PDS) formatted Voyager Experiment Data Record (EDR) files. The output cube will contain tranlated labels in pvl format with instrument, archive, bandbin, kernel, and nominal reseau information.

But it doesn't work with IMG files?

CODE
$ voy2isis from=C1625003_CALIB.LBL to=foo.cub
**USER ERROR** Input file [C1625003_CALIB.LBL] does not appear to be a Voyager EDR.

$ voy2isis from=C1625003_CALIB.IMG to=foo.cub
**USER ERROR** Unable to set PDS file.  Decompressed input file [C1625003_CALIB.IMG] does not appear to be a PDS product.


So the IMQ files are the older versions and work with ISIS, while the IMG/LBL files are newer, but they don't work with ISIS?

I tried adding
MISSION_NAME = "VOYAGER 1"
into the LBL file but pds2isis didn't pick it up.

So I tried adding it to the .cub file, but got the same error with spiceinit.

So is there any way to edit the LBL or cube file to get spiceinit to work with the IMG files? I've got a pipeline set up to use the PDS archives so would prefer to use the IMG files if possible.
Go to the top of the page
 
+Quote Post
JohnVV
post Sep 3 2016, 07:53 PM
Post #10


Member
***

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



this will take some time
first "voy2isis" takes as input the Gz ( Gnu zip) IMQ file and not the IMG

you would need to use pds2isis or vicar2isis

and manually ad the header data using editlab or similar tools

or

use vicar ( it is on github ) nasa released the source
https://github.com/nasa/VICAR
-- i have NOT used this on voyager img/lbl's


give me a few days to see what i can do

--- from just this afternoon and early evening
there will be NO!!! "quick fix"

so

you will need to manually copy out the header in the img and save it
then edit that header to convert it from the pds2 sudf format to the pds3 format
then use pds2isis
then MANUALLY add sections to the cub label using editlab
and all before you can run "spiceinit"

it is WAY EASIER to use the imq compressed pds images using "voy2isis" than the uncompressed img files and "pds2isis"
Go to the top of the page
 
+Quote Post
Brian Burns
post Sep 3 2016, 11:28 PM
Post #11


Junior Member
**

Group: Members
Posts: 54
Joined: 7-July 16
From: Austin, Texas
Member No.: 7991



(I just missed your edit - looks like we came to the same conclusion!)

Thank you - I started looking at some of the source code though and it looks like voy2isis does a lot of things to the resulting cube file, so I should probably use that, which means I should use the IMQ files. It looks like they have the LBL information embedded in them - it's too bad the IMG/LBL files aren't handled - I guess people don't use Voyager images with ISIS very much? Anyway, it's something I could file a feature request for, and depending on how long it might take I could possibly switch to using the IMQ files.

I downloaded a few IMQ files from here - http://pds-rings.seti.org/vol/VG_0xxx/VG_0...PITER/C1620XXX/ and was able to ingest them and attach the SPICE data and project a grid onto it - very cool, very simple! (aside from the C kernel pointing problems)

$ voy2isis from=C1620000.imq to=im.cub
$ spiceinit from=im.cub web=yes
$ grid from=im.cub to=imgrid.cub
$ qview imgrid.cub

Attached Image


So I think I'll file a feature request on the ISIS site - and in the meantime I'll see if I can make some mosaicks / maps with the IMQ files. smile.gif
Go to the top of the page
 
+Quote Post
Brian Burns
post Sep 4 2016, 04:18 PM
Post #12


Junior Member
**

Group: Members
Posts: 54
Joined: 7-July 16
From: Austin, Texas
Member No.: 7991



So here are some steps to make a mosaic map (might be better ways to do this) -

First download some IMQ files, e.g. from http://pds-rings.seti.org/vol/VG_0xxx/VG_0...PITER/C1620XXX/

# convert the IMQ files into ISIS image cubes
$ voy2isis from=C1620000.imq to=im00.cub
$ voy2isis from=C1620004.imq to=im04.cub
$ voy2isis from=C1620008.imq to=im08.cub
$ voy2isis from=C1620012.imq to=im12.cub
$ voy2isis from=C1620016.imq to=im16.cub
$ voy2isis from=C1620020.imq to=im20.cub

# attach SPICE position and pointing information to each cube, from the web
$ spiceinit from=im00.cub web=yes
$ spiceinit from=im04.cub web=yes
$ spiceinit from=im08.cub web=yes
$ spiceinit from=im12.cub web=yes
$ spiceinit from=im16.cub web=yes
$ spiceinit from=im20.cub web=yes

# initialize a mapping file
$ maptemplate map=map.txt projection=SimpleCylindrical clon=0.0

# project the first image cube to a simple cylindrical map, and save the updated mapping file
$ cam2map from=im00.cub to=map00.cub map=map.txt > map.txt

# project remaining image cubes to simple cylindrical maps using the same parameters
$ cam2map from=im04.cub to=map04.cub map=map.txt pixres=map
$ cam2map from=im08.cub to=map08.cub map=map.txt pixres=map
$ cam2map from=im12.cub to=map12.cub map=map.txt pixres=map
$ cam2map from=im16.cub to=map16.cub map=map.txt pixres=map
$ cam2map from=im20.cub to=map20.cub map=map.txt pixres=map

# now combine all the maps into a single mosaic map
$ ls -1 map??.cub > map.lis
$ automos fromlist=map.lis mosaic=mos.cub

# view the mosaic
$ qview mos.cub

# convert it to a jpeg (3.5MB)
$ isis2std from=mos.cub to=mos.jpg format=jpeg

Attached Image


There are better ways to combine the maps so the seams don't show so much, using averaging, or the program noseam, but I haven't tried those yet. And there are pointing errors, of course...

There's a program jigsaw, which seems like it might help with correcting pointing information, given correspondences between images (which can be calculated using findfeatures)-

findfeatures applies feature-based matching (FBM) algorihms using the full suite of OpenCV feature matching frameworks. The points detected by these algorithms are associated with special surface features identified by the type of detector algorithm designed to identify certain charcteristics.
https://isis.astrogeology.usgs.gov/Applicat...ndfeatures.html

jigsaw performs a bundle adjustment on a group of overlapping Isis 3, level 1, cubes from framing and/or line-scan cameras. The adjustment simultaneously refines the selected image geometry information (camera pointing, spacecraft position) and control point coordinates (x,y,z or lat,lon,radius) to reduce boundary mismatches in mosaics of the images.
https://isis.astrogeology.usgs.gov/Applicat...saw/jigsaw.html

Will eventually try these out also.

One problem is that cam2map is very slow - it takes about a minute on my laptop to do a simple cylindrical projection, though I'm using a virtual machine so I'm not sure how much of that is overhead. Hopefully there will be a way to speed this up.


Go to the top of the page
 
+Quote Post
Brian Burns
post Sep 6 2016, 11:12 PM
Post #13


Junior Member
**

Group: Members
Posts: 54
Joined: 7-July 16
From: Austin, Texas
Member No.: 7991



I filed a feature request with ISIS to handle the Voyager IMG/LBL files - it looks like the labels in the IMQ files are fairly different from the LBL files though, so it might be a pretty involved change. It seems that no one has been using those with ISIS, as they came out in 2012 and I can't find any mention of the problem anywhere. https://isis.astrogeology.usgs.gov/fixit/issues/4345

So I'll need to switch over to using the IMQ archives (I guess you call those the EDR archives, and the newer ones the PDS archives?), which mainly differ in the way their directories are organized.

Here's the program qmos with some images map-projected - it lets you visualize where the images end up.

Attached Image


And once you have a good map you can project it back into camera space to basically take a picture of it, or pull back missing channel info to colorize an image.

qmos can also show control points that are supposed to align the images together - I'm still trying to figure that part out.

On a side note, working with ISIS in a Linux VM is actually pretty nice - just takes a little work to get the VM set up. It's basically like running any other program in Windows - you can Alt-Tab out of it, use the clipboard, use the files on your Windows drive, etc. I've been using Xubuntu 16.04 in VirtualBox - they only officially support Ubuntu 12.04, but so far I've been able to find the libraries to get it to work okay - there's some info here - https://isis.astrogeology.usgs.gov/fixit/issues/4330. If you want to try it, just make sure to get a 64-bit Linux and set your BIOS to allow virtualization, as ISIS is 64-bit.
Go to the top of the page
 
+Quote Post
Astroboy
post Sep 3 2017, 05:37 PM
Post #14


Junior Member
**

Group: Members
Posts: 45
Joined: 27-August 14
From: Private island on Titan
Member No.: 7250



Just in time for the 40th anniversary festivities, I quickly made some really important strides in image processing in the past week that'll help me pull off a lot of ideas that have been in my head for a long time.

Firstly, I finally figured out how to batch convert pds images to image cubes, and then batch convert the cubes to png using ISIS3 (I've never been able to try IMG2PNG since I don't have a PC or a virtual machine or anything like that), using a version of some code JohnVV sent me a long time ago that I modified for my own purposes. I used to convert pds images in ISIS to cubes one at a time, then each of those cubes one at a time to fits, and then each of those fits one at a time to .pngs with my own customized contrast selections using QFitsView. Then I would stabilize each image by hand in Final Cut Pro. This unbearably long process wasn't that bad if I just wanted to work on a few images, but I'm a video guy, so every project I really wanted to do was too daunting to pull off and then I would lose interest for a while. When you jump into an area like this with little or no context it can be very difficult and confusing to follow at first, especially when you don't come from a coding background, but with the help of people on this board, I feel much more comfortable and understand how to do the specific stuff I want to accomplish. Thank you all so much.

So my png conversion is now fully automatized, but what about the image stabilization? For a while I've known that the motion tracker on Fusion 8 - a video processing application I use a lot - is incredibly smart. But shortly before I learned how to batch convert images in ISIS, I applied Fusion's motion tracker to Voyager images and it works extremely well. Yet another back breaking activity I no longer have to do myself. Additionally I found a motion interpolation plugin for Final Cut Pro that works absolute wonders. This should work really well for creating fringeless color videos. The tests I've done combining these three techniques I'm now using are truly thrilling, and I'll be posting some results in threads for the relevant planets/encounters.

Here is my customized code for converting an entire folder of images to cubes and then pngs. For this example code I would have chosen a manual stretch that works with the calibrated calib and geomed level images, since I find those versions of the images most helpful as pngs for my purposes, but I have yet to find one manual stretch that works well on all calibrated images so I can get at least Voyager-accurate color. A stretch that looks excellent on some images causes others to either appear far too dim or have clipped whites. So, for the sake of including a reliable stretch at all, I included a stretch with 255 positive values, which represents the tones of the 8-bit raw images.

CODE
#!/bin/sh
ls *.LBL | sed s/.LBL// > cube.lis
pds2isis from=\$1.LBL to=\$1.cub -batchlist=cube.lis
ls *.cub | sed s/.cub// > cube2.lis
isis2std from=\$1.cub to=\$1.png mode=grayscale format=png quality=100 stretch=manual minimum=0 maximum=255 -batchlist=cube2.lis


Voila!


--------------------
aka the Vidiconvict
Go to the top of the page
 
+Quote Post
JohnVV
post Sep 3 2017, 06:40 PM
Post #15


Member
***

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



for bulk processing isis3 has a built in set of basic tools, as in using a list

or for more complex things you can use the bash shell and the full set of terminal tools and codding - including loops and if's

the main issues i have had is with the reseau( or however one spells that) points ( rmrex)

you also might want to run "voycal" on them
-- also Voy2( i think it is 2 , need to double check on that ) data has a bit of extra noise and there is a tool for those images

voy2isis for the IMQ data
voycal for all of it
voyramp for the needed voy2 data
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 - 02:00 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.