Introduction
Centaurus-A is one of the most iconic galaxies in the southern sky.  It's large, bright, colorful, and it's full of detail so it's a common target for imagers far enough south to see it well above the horizon.  My last effort on this galaxy was just a little over two years ago with my former CDK20 and I've never been very happy with the way that I processed that data.  So, I was ready for a rematch.

After commissioning my new 24" scope in March, I noticed that Centaurus-A was rising in the East and I thought that it might be interesting to give it another go to see if I could produce a better image.  This would be my first imaging project not connected to commissioning the new scope.  When I started this project in late March, the fall weather in Chile was still delivering excellent seeing conditions and there were a number of nights when the seeing monitor showed conditions hovering around 0.5" for periods of time.  Even when it got "worse", the conditions still rarely rose very much above 1".  This was a great opportunity to go galaxy hunting and Centaurus-A is an outstanding target.  In addition to detail, color, and interesting features, it has a lot of Ha emission features that are hard to pull out of straight RGB data.  To see that stuff, you have to take a fair about of NB Ha data.  Since the moon was up when I first pointed at this object, I started exposing through the Ha filter.  As the moon waned, I set up SGP to gather Ha data while the moon was up and RGB data after the moon set to maximize use of the  entire night.

Gathering Data
I sorted the Lum data with an upper limit of 1.35" and a lot of it was right around 1".  I was stunned to see that I had gathered a fair amount of sub-arc second data through the red filter.  In almost all of my data runs, data from the blue filter gets short shrift.  I never take as much as the other channels and because of the shorter wavelength, the data is never as good as the red channel.  So this time around, I spend an extra night gathering just blue data.  I set the upper threshold for all of the RGB data at 1.5" just so that I would have a lot of data to work with.  Nearly half of this data set (23.5 hours) is from the time I spent gathering good Ha data.  I've included two processed images, one with and one without the Ha data so that you can blink between the two to see what Ha adds.  I've also posted the raw stacked Ha image.

Processing Details-Sorting and Stacking
I've been settling in on a new processing work flow--mostly based on trying to zero in on better color results.  Before I start processing, I calibrate all of my data using master dark, bias, and flat files.  then I use SubFrameSelector to sort by FWHM and Eccentricity.  I then pick the best image through the red filter with a balance of low FWHM and Eccentricity to use as a master image for alignment.  Then, I align all of the red data to that master using the rigid linear alignment option and stack it using the linear fit option for image rejection (this is most appropriate with large data sets).  I then use BXT to sharpen the stacked red data and use that result as a master alignment frame for the other channels.  Here I only need to sharpen just the stars but it doesn't matter if everything gets sharpened.  The idea is to provide less uncertainty in the star positions in the master frame--hopefully to slightly reduce registration uncertainty.

Processing--Controlling Color
I then combine the raw R, G, and B stacked data sets to create a pure RGB image.  I plate solve that image and use SPCC to color calibrate the result.  In this case, I used the Average Spiral Galaxy as the white reference.  I'm not completely sure why,  but I have found that by applying SPCC to pure RGB data, I get a much better result than if I apply SPCC to LRGB data. After color calibration, I apply BXT to the RGB result.  Here, I've tried not to over-do the sharping.  After experimenting a bit, I settled on 70% for the non-stellar components.  And, this is where the processing splits between the straight LRGB result and the LRGB+Ha image.  Here, I'll just discuss the process of adding Ha.

Adding Ha
I use the method developed by Vicent Peris and Martin Pugh, found here:  http://www.arciereceleste.it/tutorial-pixinsight/cat-tutorial-eng/85-enhance-galaxy-ha-eng to add pure Ha signal to the RGB image.  This is an outstanding and pretty easy way to add only Ha to the underlying image and to get the colors right.  One issue not addressed in the description of the method, revolves around the residual background level of the Ha data after the scaled red signal is subtracted.  It is NEVER zero--and that residual will add a faint red cast to the background of the final image.  So, I used pixel math to remove the median from the pure Ha signal and that helps quite a bit, but in most cases, it's still not sufficient.  So, I resorted to first applying a sigma filter with the MorphologicalTransform and then CloneStamping out the small residual background noise and remaining star residuals to zero values being careful to not touch the actual Ha signal.  This completely eliminates any errors in the image outside of where there is pure Ha signal.  The only other trick is to add in the "correct" amount of Ha on top of the RGB image.  Adding too much produces glaring bright magenta/red Ha features that doesn't look natural.  This is very common error that I've made in the past and that I see in other images here on AB.  It only take a little bit of Ha to look "right"!  Creating an RGB image with super super bright, saturated magenta Ha regions messes up the overall visual impression of the underlying color balance and looks cartoonish (IMHO).   Zoom way in on this image and you'll see a lot of complex Ha nebulas all throughout the central band.

Stretching and Combining
I've been using ArcSinhStretch to stretch the RGB data to better preserve color information in the bright regions.  The result requires more stretching to get to an acceptable image and I typically use the CurvesTransformation tool and the HistogramTransformation tools to get the stretch where I want it.  I just use the stretch parameter provided by the ScreenTransformation tool to stretch the Lum+Ha data.  Once I have stretched both images where I want them, I use the ChannelExtraction tool to create the R, G, and B channels that I need to use the LRGBCombination tool to put it all back together as a final LRGB sharpened and stretched image.

Final Touch Up
At the end, I still wasn't quite happy with the color balance.  SPCC produced a result that looked a bit too yellow with a very slight greenish cast so I shifted the color balance a bit away from yellow and more toward blue to produce a slightly more white galaxy dome.  As usual, I add the finishing touches in PhotoShop.  That's where I typically set the correct black level (I shoot for a value of 15) and fine tune the final color balance and saturation levels.  I also use PS to flip the image (to undo the mirror image produced by the ONAG) and to crop to the final size.  In this case, I looked at the sharpness of the image to decide if I wanted to resample it.  As I've said in the past, this scope is pretty seriously oversampled with 3.76 micron pixels at 0.18"/pixel.  So I looked at the FWHM in the final result.  Using the DynamicPSF tool, I measured the FWHM to be in the range of 3.5 -3.6 pixels with a Moffat order of 2.19.  That's a slightly broad PSF with a FWHM of 0.63" to 0.65", which is pretty good!  I estimate that I could reduce the sampling by about 75% and still maintain good image fidelity but I left this image at the native 3.75 micron sampling interval.

Interesting Features
Centaurus-A has a lot of interesting detail in the equatorial band so be sure to zoom in to take a look.  The center of the galaxy contains an active supermassive black hole with a mass estimated to be around 55 million suns.  It emits relativistic jets that collide with interstellar material to produce a strong source of both X-rays (very short wavelength) and radio emissions (very long wavelength).  One of the inner jets can be seen in this image above and to the left of the core.   It shows both blue and Ha emissions.  The X-ray jets have been measured at a length of thousands of ly long and the radio jets are over a million ly long.  Centaurus A is one of the brightest radio sources in the sky and if you include both radio and X-ray data, it subtends almost 8 degrees!  Centaurus A’s is thought to be the survivor of a collision between two galaxies that began more than 100 million years ago.  The resulting giant elliptical galaxy contains a star nursery along the central dust ring.  The sprinkling of blue stars along the edge of the lane are young, very hot super-giant stars.

So, there you have it.  I don't normally write much about my processing steps but there it is.  Hopefully there's something in there that's useful to some of you.  Let me know if you have questions or comments.  Feedback is always welcome.

John