Project

General

Profile

Bug #5264

Reseau mark removal in Voyager images

Added by Marty Valenti about 1 month ago. Updated 5 days ago.

Status:
Acknowledged
Priority:
Normal
Assignee:
-
Category:
Applications
Target version:
Impact:
Software Version:
Test Reviewer:

Description

I have some Voyager images I'm working with, and I'm trying to remove the reseau marks. I was wondering if I could get some advice on the best way to go about doing this? With just using the remrx command I got some weird results. So, I had remrx set the pixels to null, and I used the noisefilter to try and fill these in. The result are better but the stripes in the corners are not filled in. Here are the series of commands and the resulting images:

voy2isis from=C2064910.IMQ to=C2064910.cub
spiceinit from=C2064910.cub
findrx from=C2064910_cal.cub
remrx from=C2064910_cal.cub to=C2064910_remrx.cub

voy2isis from=C2064910.IMQ to=C2064910.cub
spiceinit from=C2064910.cub
findrx from=C2064910_cal.cub
remrx from=C2064910_cal.cub to=C2064910_remrx.cub action=null sdim=5 ldim=5
noisefilter from=C2064910_remrx.cub to=C2064910_filtered.cub toldef=stddev tolmin=3 tolmax=2 samples=7 lines=7 null=yes replace=average

C2064910_remrx.png View (628 KB) Marty Valenti, 2017-12-22 10:40 PM

C2064910_noisefilter.png View (628 KB) Marty Valenti, 2017-12-22 10:40 PM

History

#1 Updated by Tammy Becker 29 days ago

  • Category set to Applications
  • Status changed from New to Acknowledged

Hi Marty,

We will need to look into the weird results when remrx is defaulted to fill-in.

In the meantime, running remrx to set the pixels to nulls is a preferred work around.

Rather than noisefilter, try the lowpass filter application.

To replace only the NULL values, the command would look like the following:
lowpass from=remrx-null.cub to=remrx-fill.cub samples=9 lines=9 filter=outside null=true hrs=false his=false lrs=false lis=false

My 9x9 filter size is just a guess, you might want to try filling in with a series of lowpass starting with a 3x3, then a 5x5 on the result.

#2 Updated by Tammy Becker 22 days ago

I didn't catch this earlier, but it looks like you are not running voycal for radiometric calibration and I suspect that remrx expects an I/F floating point calibrated input image.

#3 Updated by Stuart Sides 19 days ago

  • Target version set to 3.5.2 (2017-01-31 Jan)

Task, find out what the algorithm is doing.

#4 Updated by Tammy Becker 5 days ago

FYI, I have run remrx on a calibrated image (output of voycal) and still get the same weird results with the default for action=bilinear.

This should also be tested on a viking image as well.

Also available in: Atom PDF