Ground-based image reduction in support of Juno Options

[PTDonnelly]

Hello everyone,

I am a postgraduate researcher studying the climate variability of Jupiter during the Juno mission. This involves analysis and comparison of ground-based observations taken from the Very Large Telescope (VLT) in Chile and data, to be used at a later date, from Juno, with ground-based observations corresponding with perijoves of the spacecraft.

My first project has been working on reducing images taken by the VISIR mid-infrared instrument (operating range between 5 and 20 microns) on the VLT (more details found below). As a result of new AQUARIUS detector (installed 2016) on VISIR, there is a pattern that plagues all of the images. This pattern causes problems with data retrieval (getting useful information about temperature and composition of the atmosphere) and is also not very nice aesthetically and not entirely suitable for publication.

The current technique for reducing this data and removing the lines works fine aesthetically, however it uses the program GiMP. This means that although it fixes the problem, it does so by smoothing or in-painting in ways that aren't truly scientific. Pixel values will be changed and information will be introduced or lost such that it actually affects the science output from the observations.

I have provided the link to a public google drive folder containing some of the raw images as a sample of what we are dealing with, I can provide more if necessary.

From my (limited) knowledge of programming (Python and IDL) I have been able to remove the central horizontal stripe, but the vertical stripes remain (although the images attached do not show this as they are purely raw images directly from the observatory).

If anyone has experience with removing detector patterns and any pattern like this or can recommend some techniques to try it would be greatly appreciated!

ESO - VISIR instrument: http://www.eso.org/sci/facilities/paranal/...ents/visir.html

Google drive folder: https://drive.google.com/open?id=0B8_Ynti1o...M3hjS3dWal83X28

edit: fixed the links

--------------------------------------------------------------------------------------------------
Thank you,
Padraig Donnelly

DISCLAIMER: All images provided are taken from the VLT in Chile and are fully credited to the European Southern Observatory (ESO).

[Bjorn Jonsson]

As others have indicated, flat fields and/or some instrument-specific processing should be possible. The problem currently seems to be lack of information.

Without more information/data, it is necessary either to attempt to construct a flat field (as suggested above by Gerald) or somehow to digitally filter the image in an attempt to remove the artifacts. Below is an image (the same one as used above by JohnVV) that has been destriped several times using different stripe widths. This is comparable to what the ISIS program dstripe does. The destriping can result in minor loss of real details but in this case the loss of details isn't very obvious.