Basic LRGB workflow part two - Barry Wilson

Basic LRGB workflow part two

Following on from background gradient removal from the combined RGB image and Luminance image the next stage is to calibrate the colour in the RGB image.

There are two processes for coulr calibration - (i) a relatively new process called "Photometric Colour Calibration" which uses the spectral data of astronomic objects within the image frame to weight the RGB colours; and (ii) the long-standing two-step process of Background Neutralisation and Colour Calibration.

I will regularly now use both methods and assess the results of each process and either choose to use one over the other or a blend of the two.

Photometric Colour Calibration

I am still exploring the potential of this new process however it's default settings offer an effective calibration process.  For example you have many options to choose as the "White reference" though the default option produces a good result in the main.

This process executes a plate solving astrometry step at the start to establish the astronomic objects across the field.  It is critical that the co-ordinates of the image are provided as well as the FL and pixel size of the imaging optical train.  The latter two pieces of information should be readily available to you and can be typed into the respective boxes in the "Image Parameters" section.

There are two methods to acquire the image window co-ordinates either by accessing the information from a database query or from the FITS Header of the image.

"Search Coordinates": Clicking this option a separate window opens which allows an online search.  The image above shows this method and I simply typed in "M 81" to retrieve the co-ordinate data.

"Acquire from Image": By clicking this box the process will select the information stored in the FITS Header of the image.  An image acquisition programme such as Sequence Generator Pro will store the RA and DEC co-ordinates of the image in the FITS Header automatically if the target has been plate solved.  The integrated master stack will not have this data but the individual sub-exposures do.  If you open a sub-exposure frame, highlight the window and then click on "Acquire from Image" and the co-ordinate data will be automatically populated into the respective boxes.

Co-ordinates obtained from database

The next step is to place a small preview window on the image delineating a representative acrea of background sky that is free from any structure or gradient.  When this is done, click "Region of Interest" in the 'Background Neutralisation" section - in turn ensuring that Background Neutralisation is selected - and then clicking "From Preview" will open up a further selection window in which you can select the corresponding preview. Drag the triangular Instance icon onto the image

Once the process is complete, a window pops open showing two graphs.  Ideally you want the data samples to be relatively closely clustered around the linear trend line.

However, rather than study the graph, I study the resultant change to the image and evaluate whether it is producing an image that resembles my chosen references (images from sources such as Astrobin that illustrate a processed completed image).

To assess the outcome of the completed PCC process you must first click the "linked" icon on the STF tool after reseting its parameters.

Linking the channels in the STF tool to evaluate the result

Background Neutralisation & Colour Calibration

This two-step process is a tried and tested method for colour calibration.  If you are carrying out both Photometric Colour Calibration and Background Neutralisation/Colour Calibration you will need to clone the RGB image: edit the Image Identifier to note which is the 'PCC" window and which is the "BN_ColCal" window.

First choose a background preview bearing in mind the sasme principles as discussed above.

Select the preview as the "Reference image".  The apply the Instance to the image.

Reset the STF tool after completing the neutralisation.

Colour Calibration step

Opening the Colour Calibration process after completing the Background Neutralisation step above allows the preview to be selected as the background reference.

In most instances I do not select a separate white reference but instead use whole image as the reference.  If you opt to use a preview as the white refernce (in this image of Messier 81 it could be the galactic core for example) you will need to deselct the "Structure Detection" section.  Harry's Astro Shed tutorials explain this process very well.

Apply the Instance to the image.

Completed Background Neutralisation & Colour Calibration

After the Colour Calibration has completed click the link icon on the STF tool to evaluate the result.

PixelMath to blend the two calibrated RGB images

You can combine the two calibrated images if you want to benefit from the features produced by the separate processes.

PixelMath can achieve this simply by adding together the data from each corresponding pixel in the proportion that you specify.  In the example below I have chosen to add 50% from the "PCC" image to 50% of the "ColCal".  Other proportions can be chosen of course remembering that the total must sum to one, eg 0.2*Image_1+0.8*Image_2.  The symbol "*" means multiply.

A convenient expression that enables you to state one value and automatically calculate its inverse is (without the quotes):

Typing in the "Symbols" box "n=0.2" or whatever decimal fraction you choose.

The "~" symbol calculates the inverse and in our example this is 0.8.

Clicking the square apply icon then produces a new window of the blended RGB image.

Curves boost to the RGB image

Prior to the stretching of the linear RGB image I often provide a very small boost to the colour data by using Curves adjusting the "C" and "S" curves with a slight lift.

Masked Stretch of the linear RGB image

I have found that a Masked Stretch better prerves the colour in the RGB image than using the Histogram Transformation tool.  Optionally prior to applying the Masked Stretch process you can further enhance stellar colour by using the HSV Separation Repair script and also the ArcSinh process - see the separate tutorial on these steps, here.

Choose a small preview as a representative sample of background sky and select this in the 'Background reference" window.  I keep all other options as default.

Before applying the Instance deactivate the screen stretch.

Stretched RGB image

Removal of residual green cast with SCNR

In most instances the now non-linear stretched RGB image will have a green cast across the frame.  The degree of greenish cast will vary depending on the environmental conditions where you image, eg the amount of light polluition.

You will therefore need to experiment on the amount of SCNR to apply to the image.  From urban areas in the UK I imagine that a value of "1.00" (the default) will be necessary.

If you have the Histogram Transformation tool open and tracking the changes to the window (by selecting the "tick' icon) you will be able to see the effect on the histogram of applying the chosen SCNR value.

Adjusting the histogram

It is important to note the details in the histogram after the stretch as you ideally want to align the lefthand edge of each RGB channel in the histogram.

Above you can see that the Blue channel is slightly off-set to the right and we want to better align it to the Red and Green channels.

Firstly open the preview window by clicking on the circle icon.  The click on the blue channel by clickin on the blue square.

Adjusting the Blue histogram

The lefthand triangual marker is now dragged fractionally to the right and this moves the blue hsitogram fractionally to the left.

By watching the upper window of the entire histogram you can watch the individual Red, Green and Blue histograms align.

In the previw window to the left you can see the effect of the histogram adjustment.  The background sky changes from its mild blue cast to a more uniform dark grey of the background sky.

With the previw window open, click the square apply button and the histogram transformation is applied to the image.

You can now minimise the stretched RGB image and work on the Luminance image.