Image Processing on Astro Pixel Processor
Hi Everyone, I know there's plenty of tutorials on using PI and Photoshop, but another awesome tool to process things and get you most of the way there is Astro Pixel Processor.
It's effortless to use narrowband, mono, and OSC with this program, and while you may want to (and I always do) use another program like Photoshop to tweak things, many would be more than satisfied to operate what it spits out. For this, I am focusing on pre-calibrated images from telescope.live since it is a great source for remote imaging and has access to loads of awesome data from the one-click observations and pro datasets.
When you open it, you will have to pick a working directory, so I always pick the folder that I have all the images from whatever dataset I am working on stored in. As you can see, I have already named my DSO name since I had already processed images earlier and cleared that out. It is important to remember that you have to give an object name or
APP will not complete integration. Once you select the folder you're working from, hit open.
Next we will focus on tab one, which is where you load your data in. You will want to click on light for any of your lights. One thing to note is that APP does NOT allow the use of .xisf format that is natively used in PI, so if you were to get data in that format it will need to be converted to another such as .fits or .tiff before you can use it.
One more thing to note is while this can use .jpg images, I wouldn’t recommend it at all. It does take most raw formats from various manufacturers if you’re using a DSLR, and if yours isn’t supported you can always convert them to Adobe DNG format and they will work fine.
Okay, now that we are here we’re going to load up the images that we are going to use. For mono or NB images you need to use only one type of that image since it will ask you what channel it is after you’ve selected them.
Here is where I have drilled down into the folder that contains the blue channel. Click in, select them all (keyboard shortcuts work for this) and then hit open.
Now that I have selected them I am going to select the channel they belong to.
You will get a popup asking what session it is. Since this is all from one session we will leave the only 1 session box checked and move along.
Go ahead and do the same for all of your channels. When done you will see something similar to this.
Now you can move to tab 6 to integrate everything at this time, but I wouldn’t recommend that right now because as you know there are going to be bad columns in some of the data and you want to fix that. You will also want to go after hot pixels as well.
Place a checkmark in all three spots on cosmetic correction. I find that the default hot pixel kappa of 3.0 seems to work well, but I lower the cold and hot column kappa to 6.5 and have found it works pretty well. Don’t his the create masters and assign to lights at this point. Just putting in your settings will allow it all to be done by the integration button on tab 6 at once.
On tab 3 I generally leave the star detection at 500 stars if I have nice data because it provides a more robust rejection of bad stars. So if you have eggs or donuts and you didn’t go through your subs and toss them preemptively this will catch some of them and tell you registration has failed on them and ones that is still accepts it will give less weight to them in an integration, though ideally you don’t want them in your stacks in the first place. Don’t hit the analyze stars button at this point.
Next is the register tab, and here’s where things can get fun because you can take wildly different image scales and stack them all together! Since we aren’t doing that at this point we will leave things at their default.
If you are stacking different things uncheck the same camera and optics box and depending on the scale you can set the scale stop all the way up to 15. Don’t hit the register button yet.
On tab 5 you will see a checkmark in the neutralize background tab. If dealing with lots of nebulousity in an image uncheck this, otherwise leave it checked as default. Do not hit the normalize lights button.
On tab number 6, Integration there are a few changes you can make. I generally leave the top ones at default.
If doing a mosaic or using lights from different sessions, or if using lights with different scales I would set MBB at 15% or so. If you’re looking to blend out stacking artifacts at the edges setting this at 5% or 10% is fine. It is extremely useful if doing OSC data. If you have wide variations such as from different nights, different camera and optics, or doing a session that has the moon coming up in it then LNC is a very useful thing to have turned on. A 2nd degree with 3 iterations can get a really nice result in these cases, but sometimes you might want to increase degree or iterations, but do that slowly rather than going straight for the most iterations and degree because it will make things take longer and you might not see an improvement.
Now scroll to the bottom and hit the Integrate button. This is where you will be prompted to name what you’re doing, and if you don’t the integration will cancel itself.
Okay, now the process has gone through and you should have something in your image window. Unless using an OSC astro camera or a DSLR it will be grayscale and you still need to do further steps.
Now on to tab 9, which is where we will work on a color image. You want to select the combine RGB button.
Now that you’ve done this you will click on the add channel button. It will bring a popup window that is going to start out in the last place you pulled your lights from. In my case it was on the red channel. You will want to go up to the main working directory you had selected earlier. In it you will see your master files that have been stacked. Select them all.
It should detect what kind of files each are, but will still ask you to verify what they are.
After you do this hit the (re-)calculate button. If using luminance files you will notice in many cases your image will look kind of washed out.
You can click the saturation button on the processing pane and it will being some back, but in some cases you will want to bring the luminance slider down, or with some globs that are really bright simply remove that channel with the remove button up above where your sliders are.
In the first instance I have lowered the luminance multiplier from 1.000 to 0.192
Next hit save and then name it whatever you want to, or simply leave it as the default. Make sure you’re using the save in the upper left of the window, not the save on the processing pane toward the right. That way you can try different things with the sliders and save each iteration, but you should name them something like object_v2.fits
When done hit the cancel button up there to return to the main processing screen. It will take you back to the tab with tools selected, but will not have an image in the viewer, so select the one that is labeled as Other/Processed down at the bottom since that will be your image with the color channels combined. Double click and it will come up and be loaded. It will also be highlighted down below.
At this point you can use the stretch presets next to the red question mark and the saturation, sharpening, etc. buttons for processing. You can also leave it with no stretch too. You can even do a inversion on the data which is where you see a white background and objects are dark.
When you get it to where you want to save it you will use the save button over in the processing pane. At this point you can save it as a fits file, jpeg file, or tiff file. If you wanted to tweak in another astro program that uses fits, save it as that. For Photoshop I would save it as a 16 bit tif file, which you will get those choices when doing your saves. Do not pick 8 bit because that has less data in it, and do not pick 32 bit because it will blank out some functions in Photoshop. If you’re happy with what you have and think it is in a state you’d publish it in then save it as a jpeg.
A few more functions that I didn’t touch on are the other tools in there.
The batch modify will allow you to split the channels of RGB images so you can stack in mono images, useful if you were to combine data from different sources.
The remove light pollution one is great if you have any weird gradients, but take note, it can also remove some extended nebulousity in some cases. It is very useful and will give you a very nice background and will only work on what it sees as the background. When making selection boxes for it always select areas that are background and not stars. Then hit calculate, and if you like the result hit OK & SAVE. You can keep working this if your background doesn’t even out by adding more boxes and recalculating it.
The same goes for calibrate background as it does for remove light pollution.
You can only do calibrate star colors if you’ve either ran the remove light pollution or calibrate background tools. Sometimes this gives a nice result, sometimes it gives oen that to me messes up your color balance so I will hit cancel on it. Note that it will also change the color balance of everything in the image, not just the stars. You want to select only one box for it to use.
HSL selective color allows you to change your color balance and tweak other post processing settings on the image as a part of the process. It works on a per channel basis.
The rest should be fairly self explanatory based on the names. I do not use the correct vignetting since I can get a better result with the remove light pollution tool, and I only do my resizing and rotating after I have processed the image to my tastes.
So there you go. This is a quick rundown of how easy APP is for image processing. One more nice thing is if you decided to work another image all you have to do is change the directory by clicking on the button that shows where your working folder is, then hit clear to clear out all data from the previous object, then load new stuff in, then go to the integrate tab and hit integrate.