How can I remove banding effect from .cub file Magellan Venus Options

[ncc1701d]

I have this .cub file of a section of Venus taken by Magellan that I got from map-a-planet2. Picture included below. You can see the striping effect. Does anyone know of a way to remove this banding effect easily in ISIS3 or another tool?
The best answer I have to far is using "Equalizer" in ISIS3
https://isis.astrogeology.usgs.gov/Applicat.../equalizer.html


Problem though is it looks like you have to fix it using the original striped input images before they are assembled into a .cub.
All I have though is the .cub that was already mosaicked.
Any suggestions on how to remove the striped effect at this point?
I dont believe those orginal individual striped images are avaible to public.
All I know of is the 130 gb .cub file found on mapaplanet2 from which this image was taken from.
thanks

Attached thumbnail(s)

 

[Bjorn Jonsson]

dstripe in ISIS3 can remove striping but the problem is that the stripes have to be vertical or horizontal (maybe you could rotate the image before destriping?).

See https://isis.astrogeology.usgs.gov/Applicat...pe/dstripe.html

[mcaplinger]

I dont believe those orginal individual striped images are avaible to public.

Sure they are. They're called Full Resolution Basic Image Data Records or F-BIDRs. See http://pds-geosciences.wustl.edu/missions/magellan/

Getting rid of shading at the boundaries of the swaths was extremely difficult because of radar artifacts. Maybe in the 20+ years since I've looked at Magellan data someone has improved the processing.

[JohnVV]

if you want i just gave a copy of my 131072x65536 pixel de striped Venus map to the Mesa Arizona CC planetarium
from the C1 data set

this will save you about 6 months or more of work
the black and white is 8 Gig the color is 24 gig in isis3 cub format and georeferenced
--

--





Dsrtipe is what i used after remapping to Mercator and rotating sections of the C1 data to vertical
( the strips are on a " s "curve )

for the missing "No-data" areas i inpainted using the Resynthesizer plugin for gimp
combined with inpainting using the Greyc laboratories "Greystoration" tool

I dont believe those orginal individual striped images are avaible to public.

yes they are
the C1 is here
http://pds-imaging.jpl.nasa.gov/data/mgn-v...5-midr-c1-v1.0/
the FULLRES "F-BIDR"
http://pds-imaging.jpl.nasa.gov/data/mgn-v...-full-res-v1.0/

[JohnVV]

an example from the C1
http://pds-imaging.jpl.nasa.gov/data/mgn-v...0n335/c1f12.img
http://pds-imaging.jpl.nasa.gov/data/mgn-v...0n335/c1f12.lbl

CODE

pds2isis from=c1f12.lbl image=c1f12.img to=c1f12.cub

use "qview's" tool measure for the angle
-- 5.6 in this case

CODE

rotate from=c1f12.cub to=work.cub degrees=5.6

CODE

dstripe from=work.cub to=work1.cub vlnl=137 vlns=3 vhnl=3 vhns=137

CODE

rotate from=work1.cub to=c1f12A.cub degrees=-5.6

( you will need to fix the mapping and extra border on the image after the second rotate

[Ian R]

That's amazing, John. Somewhat on a tangent, could these techniques be utilized to clean-up the USGS Titan radar mosaic, as seen here:

http://astrogeology.usgs.gov/search/detail...n180_128ppd/cub

Furthermore, as someone who doesn't have access to ISIS, would the CUB file preserve the SAR swathes in separate layers?

[JohnVV]

ian

for that cub ?? no

if you work with each of the radar passes and mosaic then one at a time ?? then maybe ??
but "noseam" would be a better tool

back on target for the thread

The dstripe tool works best on very regular noise and for it's intended use
-- see the images here
https://isis.astrogeology.usgs.gov/Applicat...pe/dstripe.html

[mcaplinger]

Given how well this worked, I'm surprised no one has gone back and reprocessed the whole C1-MIDR dataset like this. There's this -- http://www.celestiamotherlode.net/catalog/...p?addon_id=1070 -- but I haven't looked at it.

[JohnVV]

mcaplinger

that IS my map on the Celestia motherload
( libido also ask me not to host any more 12+ gig files )

in 2011 / 2012 over the winter i reworked the C1 data

[mcaplinger]

in 2011 / 2012 over the winter i reworked the C1 data

Is the georeferenced CUB version online anywhere? Maybe USGS should be using that since it's so much better than the original version.

[JohnVV]

my large map is here:
-- do to age the below links do not work
if someone needs it PM me
https://drive.google.com/file/d/0B6ZYAd08tZ...iew?usp=sharing
and
https://drive.google.com/open?id=0B6ZYAd08t...SXpvS1V1T3R5bDg

this 4.91 gig zip ( would be smaller if i used xz archive) unzips to a 8 gig , 1 band isis3 cub file 131072 x 65536 pixels


this is the header

CODE

Object = IsisCube
  Object = Core
    StartByte   = 65537
    Format      = Tile
    TileSamples = 2048
    TileLines   = 2048

    Group = Dimensions
      Samples = 131072
      Lines   = 65536
      Bands   = 1
    End_Group

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

  Group = Mapping
    ProjectionName     = SimpleCylindrical
    CenterLongitude    = 0.0 <degrees>
    TargetName         = Venus
    EquatorialRadius   = 1737400.0 <meters>
    PolarRadius        = 1737400.0 <meters>
    LatitudeType       = Planetocentric
    LongitudeDirection = PositiveEast
    LongitudeDomain    = 180 <degrees>
    MinimumLatitude    = -90.0 <degrees>
    MaximumLatitude    = 90.0 <degrees>
    MinimumLongitude   = -180.0 <degrees>
    MaximumLongitude   = 180.0 <degrees>
    UpperLeftCornerX   = -5458203.0763469 <meters>
    UpperLeftCornerY   = 2729101.5381735 <meters>
    PixelResolution    = 83.285569402241 <meters>

    # PixelResolution    =  <meters/pixel>
    Scale              = 364.088888889 <pixels/degree>
  End_Group
End_Object

Object = Label
  Bytes = 65536
End_Object

Object = History
  Name      = IsisCube
  StartByte = 8590000129
  Bytes     = 1748
End_Object
End