Calibration & Assembly - Barry Wilson

Calibration & Assembly

This tutorial sets out the steps I use to calibrate my images and assemble them inot integrated stacks for each channel.  I use these steps whether processing an LRGB or NB image.  There are additional steps required for a mosaic and I plan to cover these in a future tutorial.

The tutorial assumes that you have already a library of master darks, flats and a bias correlated to your imaging temperature and sub-length as appropriate.  I will briefly describe how to create a master calibration file below if you have not yet completed this step.

The steps I take are a mixture of the script for Batch Pre-Processing (BPP) and manual assembly using the Star Alignment and Image Integration processes.  You can use the BPP script for the entirety of calibration and assembly of master stacks.  It uses the Star Alignment process as part of its automation however using it manually enables you to select some further options.

The same applies with BPP's integration: whilst it is ideal for calibration file masters, ie your darks, bias and flats, manually using Image Integration enables further options to be chosen.  The master output files are suitable for a quick test image though to evaluate the data gathered to date.

The steps I take are:

1. Cosmetic Correction - a preparatory step taken in order to save the process for use in BPP. This process helps remove the random hot/cold pixels not otherwise identified by the master bias, including column defects.

2. Batch Pre-Processing - automation of application of darks, bias and flats to remove noise and optical halation (vignetting).

3. Star Alignment - aligning all of the sub-frames to a master, the preparatory step necessary to integrate the individual sub exposures.

4. Image Integration - combing the individual sub exposures into a master stack and in the process removing artefacys such as satellite trails.

Cosmetic Correction

First, open up a sub-exposure.  In the example for this tutorial I am using a Ha sub of 1200s duration.  If you data set includes subs of differing exposure you will need to carry out a separate Cosmetic Correction process for each sub duration.  The reason is that the BPP script only has the facility to use a single saved Cosmetic Correction process.

For example, if you have data for an HaRGB image with RGB subs exposures for 600s and Ha exposures for 1200s, you will need to perform two Cosmetic Correction processes: one for the 600s exposures and one for the 1200s exposures.  The criterion here is exposure length not filter type as you seeking pixel-by-pixel deviations which are related to temperature and not filiter-type (the use of which selects photons with certain wavelength ranges).

With the sub open, open the Cosmetic Correction process.  Select "Use Master Dark" and navigate to the respective master dark file by clicking on the folder icon.

Next, select "Use Auto Defect", then select "Hot Sigma" and 'Cold Sigma".  Click on the preview icon and the preview window will open and take a few seconds to execute to identify the hot and cold pixels.  When complete (the preview icon will stop whirring) save the process by dragging the New Instance icon to the workspace.

If you have a bad column you can select its position under the "Use Auto defect" options.  By updating your bias library and dithering when imaging you will reduce the need to use the defect options.

If you need to complete a further Cosmetic Correction process for subs of a different length, you may want to amend the identifier of the saved process to something meaningful, eg "CC_1200s" to help you identify the correct process when you select it within the BPP script.

Batch Pre-Processing script

The BPP script opens presenting four tabs on the left, each for the files under the relevant tab headings.  The buttons across the bottom open up a window browser enabling you to navigate to the respective file location.

I normally select "+ Add Files" and then navigate to my calibration files folder and select the master bias, dark and flats.

Adding files

After adding calibration files you then add your Lights by clicking on "+ Add Lights".

Adding Lights

At this stage, youalso need to select the Cosmetic Correction process by selecting "Apply" under Cosmetic Correction and then using the drop down list.  Under "Global Options" you also select "Use master bias", "Use master dark" and "Use master flat".

If you do not have masters because you are integrating your individual bias, dark or flat subs for the first time to create a master do not select these options.  Instead you will need to add the entirety of the bias, dark and flat subs in the relevant tab by clicking on the respective "+ Add [name]" button and making sure the "Calibrate only" option is unselected so master files are produced.  These will be located in a "Masters" folder in the location you select.

In the "Output Directory" use the folder icon to browse to your chosen location.

I am only performing an automated calibration process as I have the "Calibrate only" option selected so I therefore do not select a specific sub to be used as the reference for star alignment (this is a separate manual step in my workflow).

Clicking the "Diagnostics" button on the bottom will display if you have any setting or option incorrectly selected.

Press "Run" and the script will execute.

When complete it will produce a folder called "calibrated" in the chosen location, within this folder is another folder labelled "light" within which are the calibrated subs each file name ending with "_c".

Also within the folder "light" is another folder called "cosmetized" with each file name ending in "_c_cc".  It is these "costmetized" files that we need for the next star alignment stage.

Repeat as required for filters of different sub length.

Star Alignment

Above you can see the Star Alignment process window complete with the necessary options and files.

The first task is to select a cosmetized sub to act as the reference frame for alignment of every other sub.  When processing an LRGB or RGB image I will use a Red frame as the stars through this filter tend to be the tightest and so giving less ability for an alignment error of their centroids.  If creating a NB or HaRGB or LHaRGB image I usea n Ha sub.

Using the arrow to the right of "Reference image" opens a window browser so you can select the reference image.

Selecting reference sub

A good tip when acquiring the data to minimise atmospheric distortions is to acquire your data when the target is at its highest altitude thus minimising the depth of atmosphere through which the photons have to travel before arriving at your scope and CCD.  Blue wavelengths are especially subject to refraction by the atmosphere so try to ensure that these subs are acquired near to the target's meridian.

To aid minimising any distortion within your data arising from the translation of the photons from a curved sky through your optics and onto the flat plane of the CCD and the atmospheric effects along the way, I tick the "Distortion correction" option.  If you have a particularly large FOV or are creating a mosaic this option helps.  I leave it on the default option for "Registration model" (I do modify this for moasics however).

Select the subs for aligning by clicking on "Add Files".  This opens a window browser.

If the image is a single frame you can align all of the subs regardless of the filter in a single application of Star Alignment.

Selecting file location for aligned subs

Under the section 'Output Images" I click the folder icon which opens a browser window.  Under my chosen target folder, I opt to create a new folder which I call "registered" and then select this folder as the location for the aligned 'registered' images.

Output images will have a file name that ends in "_r" and will be placed in the chosen location.

Click the circular global apply icon and the process will execute.

Image Integration or 'stacking'

The final step is to stack the registered aligned subs using the Image Integration process.

You select your files by clicking "Add Files".

I then select the Pixel Rejection options.  I always start with the defualt settings after selecting the algorithm under "Pixel rejection (1)", in this instance my preferred algorithm "Winsorized Sigma Clipping".  This works very well for large data sets.  The option "Sigma Clipping" works well for smaller data sets.

I then click the circular global apply icon.  The process will then output three windows: the integration and the low and high sigma clip windows which display which pixels have been rejected from the "integration" window.

If you have prominent satellite trails these will be shown in the high sigma window.  A good practise to aid in rejection of these image aberrations is to use "dithering" when acquiring the data.  If the entire satellite trail(s) is not rejected  and are visible in the "integration" window, you can amend the sigma clipping values, "low" and "high" under "Pixel Rejection (2)".  Adjust to suit your data and ensure rejection of unwanted artefacts.  Delete the three image windows and re-click the circular global apply icon after naking your adjustments.

PixInsight has also introduced a "Large-Scale Pixel Rejection" tool which I have not found the need to use.  An explanation can be found for this tool on the various web resources.

Integrated or stacked master Ha image

Once you have the master stacked image you can then save it and proceed with the rest of the processing.