Keith B Quattrocchi, MD, PhD December 10, 2005
WORKING WITH MIRA AP-7.25
(A simplified approach)
B) Guide Summary
C) Guide in 6 Steps
A) Introduction: Let’s begin with an accurate disclaimer. I am no expert on the use of MIRA AP. I am an amateur astronomer, end of story. I’ve put this manual together for one purpose. That purpose is to help those, who are without a solid background in astronomical imaging, in getting MIRA “up and running”. I’ve done this as a number of amateur imagers have asked me about using this program. They’ve heard it is a remarkable program, but need a little help with the basics of how to set things up for calibrating (subtracting out dark/bias/flat frames) and registration (aligning images). I’ve written this for MIRA 7.25, as that’s the version I’ve been using. But I’d recommend looking at Mirametrics site, as they have newer versions of this program (I believe 7.56 will be released in December of 2005) and they have new tutorial files which are quite helpful (currently samples available by clicking on “more” under MIRA AP on their homepage, and then clicking on “Image Registration” under the “Samples” section in the upper right of the page). They also have a scaled down (and incredibly inexpensive) version, MIRA AL, which may suit your needs. The website for Mirametrics is http://www.mirametrics.com/ ).
Most of what I know about this program I learned from others, and the school of hard knocks (in which I retain full time enrollment). But I have been impressed in the power of this program once one has learned to set up the “profiles” properly. As with any powerful program, it has a steep learning curve (I’m hoping this tutorial will help with that) but is EASY to use once you’ve got it up and running. The way I will recommend it you “set it up” is just one possible way to use the program. I’ll walk you through MIRA from setting up your “files”, to calibrating (dark/bias/flat subtractions), to “registration” (image alignment) and finally, image combining. To those who have not used programs like this before, it may or may not seem intuitive. But once you’ve learned to use it, it becomes quite intuitive. In addition to using this program, I’d recommend learning to use a variety of programs. I often use CCDAutoPilot III, MIRA AP, CCDStack, Maxim DL, AIP, CCDSharp, Adobe CS and others. I believe each of these programs has its’ own inherent strengths, which you will want to take advantage of.
For those of us who have MIRA AP-7, I thought a “jump start” guide would be helpful. It is a powerful program developed for astronomers and it has hundreds of features I won’t even mention (I’ll try to point out a few as I go along). The new help files are excellent, especially once you have the basics down. When using them refer to the header on the dialogue box when looking up anything in the dialogue box (you may not find it listed any other way).
In using this guide, I’d recommend you start by set EVERYTHING up exactly as I do here, at least during your first run.. That includes the names of EVERY file I list. If ANYHING is labeled incorrectly in MIRA you’ll get an error message (as with any program). So first do it my way, then you can set it up the “right” way.
Required Homework Before We Get Started: Before you can use this guide, you’ll need to take a series of dark/bias and flat frames and store them into a specific folder on your desktop. I’ll explain exactly how you can do this automatically (with CCDAutoPilot, which I strongly recommend, and can be found at http://www.ccdware.com/ ) or manually (both discussed in Addendum 1). You MUST complete this step, as outlined in Addendum 1, otherwise you’ll have NO DATA from which MIRA can make your Master Dark/Bias/Flat Frames. As I’ll discuss, for the purposes of this first run through, my critical camera parameters are -15 deg C, 900 sec images and 1x1 binning. Remember, you can change all this once you have things working and understand the program. For those who do not understand why we calibrate, and the basics behind it, I’d recommend reading up on the subject. Ron Wodowski’s book “The new CCD Astronomy” is an excellent and very readable resource (and can be found at http://www.newastro.com/newastro/default.asp ).
A Point Worth Noting Before We Get Started: I want to EXPLAIN something about how MIRA works, but in doing so I may confuse you. So you can read through this paragraph if you want to or ignore it and just follow the instructions below (the point will still become clear as you go along). I just wanted to put this in here as it’s a fundamental part of how this program works. So here it is. In using MIRA AP7 it’s important to understand the programs ability to “Grab” information which you’ve stored, such as your dark frames, and use this data. Computer folks use non-intuitive words to describe this, so I thought I’d point this out now and then later discuss it in more detail. In order to grab data the program needs to know a WHERE it is and what its’ CALLED (named). In MIRA the way to ask the program to grab something is to use a “Filter”. I guess the analogy is you’re filtering out what you don’t want and therefore keeping the data you need. Let’s say you want the program to use your 20 dark frames and subtract them from your raw data. In that part of the program (we’ll get to this later) you’ll be checking a box called “Use Filter KWD” (‘KWD’ is Keyword). You’re telling the program that you’ll give it a “keyword” (the first part of a files name), and it should grab all the data with that keyword (you’ll also tell MIRA where the data is). In another part of the program, as we’ll see, you’ll tell MIRA the name of these files (called “Source Files”). So here’s how it works. if you have a bunch of dark files, they might be called dark1, dark2, dark3….dark20. They will be in a file with other things (which you won’t want the program to use during dark subtraction). You would then use the term “dark*.*” as the “Source File” name (after telling MIRA you’ll be using a “Keyword”), and the program will grab everything in the folder that begins with “dark” (the *.* after the word “dark” tells the computer to grab everything in a specified folder which begins with “dark”). The point is, you’ll only need to name the beginning of the file properly, and MIRA will grab only those files which start with the “keyword”. OK, I’m sure you’re lost, but after you run through the steps, this will make more sense.
B) Summary to the guide:
OVERVIEW: Remember, in the end, our goal is to calibrate and align our images. First we need to do this for each filter we use (e.g., a blue filter) and THEN we need to be sure all the different filtered images are aligned. The order of work is as follows: 1) first we’ll create folders to hold our calibration data, 2) we’ll calibrate the images (dark/bias/flat subtract) for one filter, then 3) well align each of the images for that filter (i.e., all the Blues will be aligned), then 4) we’ll “Combine” all the aligned stars for a particular image (i.e., make our MASTER BLUE in this case), and finally 5) We’ll make sure each of the combined L. R, G, B, and H-alpha images (or whatever else we have) are aligned (that way we can combine them in Maxim, deconvolute in CCDStack or AIP, Process in Adobe, or use whatever programs we wish).
To do this, this process and guide is broken into 6 parts (you’ll get more detail in the next section):
I: Folder Creation: This section describes how your calibration folders should be set up. MIRA uses these folders to store its Master Frames and we use it to insert our “raw data (i.e., our dark, flat and bias frames). I’ll try to provide a little information about how to obtain these raw files, but it’s assumed the reader has done this before. The information about how to obtain these frames is in the Addendum. Again, you can not use this guide if this step is skipped (as you’ll have nothing to calibrate the images with).
II: Master Dark and Bias Frame Creation: Here I’ll describe how MIRA will make our Master Dark and Bias Frames from the raw data. The way I run MIRA, I only make the Master Dark and Bias frames when I need to (a few times each year). In contrast, the program will be recreating the Master Flat Frames every time we run the program to calibrate our images. Don’t let this bother you (as it did me), it’s just one way of using this powerful program. I do it this way as I tend to replace the Flats more often than the dark or bias frames. Therefore I let the program remake the Flats each time automatically.
III: Calibration: Here we describe how we set-up MIRA to Calibrate (subtract from our images the dark, bias and flat frame data). If you don’t know why we need to do this I’d again recommend you read Ron Wodowski’s book entitled “The New CCD Astronomy” ( http://www.wodaski.com/newastro/default.asp ). It’s an amazing book. Or if you’re able to handle a more in depth mathematical approach you can try “AIP for Windows” (which is thorough, includes the mathematics going on behind the scenes and comes with a powerful imaging program on CCD). This can be found at http://www.wvi.com/~rberry/imageprocessing/aip4win.htm .
IV: Registration: This section details how we can ALIGN our images, Why astronomers refer to this as “registration” is a mystery to me. Others call it “stacking” (which is a slightly more intuitive name for the process). What we’ll be doing is aligning our images using a very powerful alignment program. There are a few important “tricks” you need to know if you’re going to use this program.
V: Combining Images: In this section we’ll take the images we’ve calibrated and aligned (registered) and put them together as a single image. For example, we might have started with 20 separate Blue images of M74. We will first have done our bias/dark/flat corrections (Section III). We will then have aligned (registered) these 20 images so that we don’t have a blurry mess (Section IV). Now we will take the 20 calibrated and aligned Blue images and COMBINE then together. Because of the way in which we will do this (more later) this will result in increasing the SIGNAL (what we want, i.e., the data which represents the object) and decreasing the NOISE (random signals) in the image. This step will take what looks like a bunch of garbage and turn it into a nice looking image (which will look incredible after you’ve processed it further in Maxim DL, CCDStack, Adobe CS, AIP4Win and any other program you like to use). Again, Ron Wodowski’s Book goes into this in great detail.
VI: Registration of the Combined Images: Before we leave MIRA AP7 we will align the various images we’ve been working on. Typically these will be Red, Green, Blue, Luminescence and H-alpha images. By aligning them now, we can just open them up in our “Processing” programs of choice, and they will already be aligned.
C) The Guide in 6 Steps
I: Folder Creation:
First of all, in order to keep all the terminology consistent I will assume all your data is collected using 1x1 binning. Once you set up the program this way you can change to 2x2 binning or anything you want. But again, this document is a bonehead approach, and we need to keep it simple and consistent. Hence, you will see the files will all be binned 1x1. Once you’ve got the basic steps down you can change this.
The first step is to set up a calibration folder. I put this on my desktop and call it, cleverly, “All Calibration”. For now, I suggest you give it the same name (MIRA will then use this name, which we’ll later enter, when sorting information). It is simply a folder on my desktop, no more or less.
Inside this Folder we will inset a folder for our particular camera set-up. In this case we’ll use mine, and call it “Calibration STL-6303 f9 RC 15min -15deg”.
Inside this second folder create 2 new folders, one called “Darks” and one called “Flats”. Later we’ll see that the dark and bias data will end up in the “Darks” folder and the flat field data will end up in the “Flats” folder.
As well see later, MIRA will create another folder inside of the “Darks” folder and inside the “Flats” folder. This folder will be called “CALIBRATIONS”. Don’t put it in there now, this is a folder MIRA will CREATE for us. Later, inside of the “Darks” folder MIRA AP will place our MasterDark1x1.fit file, and our MasterBias1x1.fit file. Also, later, inside the our “Flats” folder MIRA AP7 will place our “MasterFlat_Blue.fit” file (also, MasterFlat_Red.fit, etc…).
For now you just need to set up the “All Calibration” folder, the folder in this called “Calibration STL-6303 f9 RC 15min -15deg” and the “Darks” and “Flats” folders within that folder. MIRA will do the rest later.
II: Master Dark and Bias Frame Creation:
Now we need to open up MIRA AP7. Once you do, go to Process/Calibrate Images. We’ll go through the dialogue box step by step. You’ll need to do the following:
1) First we will name the Profile (Figure 1, below). Simply give it a name you’ll recognize. In this case, use the name I use (again, you can change everything later). The “profile” contains all the information on what the programs doing, in this case, making a Dark and Bias Master Frame. You can use my profile name for now, and can always change it later. The small box and arrow to the right are ways to open, close, lock or rename the “Profile”. It’s worth playing with now, so that you don’t mess up your “profile” later.
2) Notice that the next line has the DOT on “Folder”. When you get the point of running the program, you may get an error message if it’s anywhere else. We are making a Master Dark and Master Bias, not calibrating our images. So for now DO NOT put the dot next to “Images”. Later, when we’re calibrating you MUST put the DOT next to “Images” (this will make more sense later….possibly).
3) Folder: This is important. You need to hit the small square to the right and navigate to the folder you set up. It should be (if you set it up as I recommended) under: C:\Documents and Settings\Owner\Desktop\All Calibration\Calibration STL-6303 f9 RC 15min -15deg\Darks. This is WHERE MIRA will find the raw data needed to make the Master Dark and Master bias frames. If you followed my advice (in the Addendum), you’ll already have raw data in this file, labeled as I’ve recommended. See Addendum 1 for this information.
4) In the upper right is a box called “setup”. In that tab click on the arrow to the right of “preferences” and check off “Save Undo” and “Overwrite Cal Frames”. The second is key if you want to add new flats (and you do). Under “Notification…” check off “popup message” and under “Create Calibrations in this Subdirectory” check “Calibrations”. Be sure to do this, if you don’t, MIRA won’t make those files in the Calibrations folder (yes, because there won’t be a calibration folder !!).
5) Now you need to click on the “Procedures” tab and tell it to “Create a Bias Frame” and “Create a Dark Frame”. The “Procedure” is what we’re doing. In this case making a Master Dark and Bias Frame. There’s lots of other confusing options, but you can ignore them for now or have a nervous breakdown trying to fathom them all out. “The choice is yours”.
6) Now it’s on to “Methods”, where we tell MIRA HOW to do what we want and what to call I (see Figure 2). Click on “Methods” and then the “Bias Corrections” tab and then “Create Bias Frame” tab (see Figure 2). Again, you can use any names you want, but I recommend you use mine for now. Under “Output Bias Frame Name” put down “MasterBias1x1.fit”. If the program defaults to an fts file name, just rename it to fit. Hit the checkmark on the right (stores this entry, and is a good habit to get into), and you might as well begin to hit the SAVE button periodically (saving the profile as you’ve made it). Do this by hitting the save button to the right of “Profile” (top of the screen). If it’s grey, go to the arrow and uncheck “locked”, then save (and re-lock if you want to play it safe). Different MIRA versions do this differently, but you’ll get the idea by messing with it.
7) Under combining method check Mean-Min/max clipped. This just means it’ll take the data and drop the maximum and minimum values at any pixel site (i.e., any noise) and give you the mean value (that’s essentially what it’s doing, the real scoop is to difficult for me to explain).
8) Now navigate to “Dark Correction/Create Dark frame” (see Figure 3 below). Now we do just what we did above, but this time for the Dark Frames. Again, I call the file MasterDark1x1.fit. Do not check “apply bias”, as this is just the Master Dark we’re making. We’ll do the bias subtraction later, when we do our calibration.
9) Now for the part I tried to explain early on under “A Point Worth Noting Before We Get Started”. Here we tell MIRA which files to grab in making the Master Darks and Master Bias frames (see Figure 4). Navigate to the “Source Files”/Bias and type in (if you’re using my terminology) “bias*.*”. Be sure the dot is on Template. Also, save the entry by hitting the checkmark (as we discussed before). Notice, again, that the “Folder” is the folder which contains these “Source Files”.
10) Now on to Figure 5. Here we do the same for the raw dark frame data. Be sure the file is named correctly. In my case the file name is “-15Dark900s1x1*.*”. See Addendum 1 for details on how to set up this file. Hit the check mark.
11) You’re done. Now just go back to the Procedures tab. Be sure you have the Profile saved and LOCKED. Then hit “RUN”.
If all goes well you’ll see a message which says “The Calibration Task has finished”. Once you close down the “Image Calibration” screen, you’ll see a window called “CCD Calibration Results”, which summarizes all the steps performed. Next go to your desktop folder called “All Calibration”, open up the folder “Calibration STL-6303 f9 RC 15min -15deg”. Under “Darks” you should see a folder (which you just made!) called “CALIBRATIONS” and in there you should see 2 files. These are your new Master Bias and Master Dark files. They should be called MaterBias1x1.fit and MasterDark1x1.fit. If they’re there, you’ve passed a major hurdle.
III: Calibration: Now we’ll do 2 things at once. Don’t let this confuse you (as it did to me !!). First, we’ll be doing our dark/bias subtraction and secondly, we’ll be making the Flats, which will be immediately used for flat subtraction (each time we calibrate in the future, we’ll be remaking the Master Flat frame). This will seem a bit odd at first, but it’ll make sense once you’ve had to re-do your Flats a few times.
So far we’ve made our Master Darks and Master Bias frames. You’d think we’d make the Master Flats next. We COULD do this, but a BETTER way (translation: the way I’ve gotten used to doing it) is to MAKE the flats each time just before we CALIBRATE (i.e., do the dark/bias/flat subtractions).
We do it this way because we frequently need to re-do our Flats (due to dust landing on the CCD imager or changes in our optical chain).
1) Now to get started, we need to open up MIRA and go to Process/Calibrate images. This is shown in Figure 6 A and 6B (I used 2 figures as some of the drop down information under “Procedures” has to be scrolled down to see). We’ll be setting the parameters so that MIRA can use the Master Dark and Bias Frames we just made, AND, so it can make and use our flats. THIS will allow us to “calibrate” our images (i.e., do our dark/bias/flat subtraction).
Important Note: We will make a profile for each filter we use, generally L, R, G, B and H-a. The Blue profile will use the Blue flats, Reds the Red Flats, etc… It’s the flats which are different for each filter (no change in the bias or darks). Remember, the order of work is as follows: 1) first we’ll calibrate the images for one filter, then 2) well align the images for each of the filters (i.e., all the Blues will be aligned), then 3) we’ll “Combine” all the aligned images for a particular filter (i.e., make our MASTER BLUE in this case), and finally 4) We’ll make sure each of the combined L. R, G, B, and H-a images are aligned to each other (that way we can combine them in Maxim, or whatever program we’re using).
Check off all the boxes as seen below. Once you read through this it will make sense. You are USING dark and bias frames AND you are MAKING flat frames in order to calibrate your image.
2) Be sure to check the “Setup” menu, and be sure it’s set up as we previously described. See Figure 7.
3) Now on to the Methods section, where we specify how the Procedure will be performed (Figure 8). We need to go to “Apply Bias Frame”. Now we need to be SURE to point the program (under “File”) to the “MasterBias1x1.fit”. You can use the button on the right and navigate there. This will be:
C:\Documents and Settings\Owner\Desktop\All Calibration\Calibration STL-6303 f9 RC 15min -15deg\Darks\CALIBRATIONS\MasterBias1x1.fit. Be sure the “dot” is on “File”.
4) Now we do the same for “Apply Dark Frame”. Be sure the file is correct. It should be:
C:\Documents and Settings\Owner\Desktop\All Calibration\Calibration STL-6303 f9 RC 15min -15deg\Darks\CALIBRATIONS\MasterDark1x1.fit.
Be sure the “dot” is on “File”. See Figure 9.
5) Now to CREATE our flat frames (Figure 10). Set things up as noted below. We want the Master Flat to be called “MasterFlat.fit”. The program will know to add the particular color filter to the name of the Master Flat, based on the “source file” information (as we’ll see). For example, the final name of our Blue flat in the “CALIBRATIONS” folder will be MasterFlat_Blue.fit. The combining method I use is Mean-Min/Max clipped. Be SURE to check “Apply Bias” and “Normalize”. For reasons I have never get quite fully grasped, we do not need to apply a dark to the flats. Check off “Use Filter Kwd” and be sure “Filter” is in the box. This will tell MIRA to use the proper “Source Files”, which is coming up.
6) See Figure 11. Now to APPLY (use) our Master Flat frames (click on “Apply Flat Frame”). Again, mark your boxes as you see below. “Result of (Create Flat Field) is telling MIRA to use the flat field we just created (not terribly intuitive to me). Again, check the filter box correctly. Notice the grayed out folder is still pointing to our FLATS file.
7) Finally on to setting up the source file (Figure 12). We only need to look at the Flat tab. The other Source Files are irrelevant, s we’re not using them at this stage (so don’t even look at them !). Using the nomenclature I set up earlier, put in “Blue_Flat0*.*” (again, this is covered in the addendum, an hopefully already taken care of). Again, the *.* tells MIRA to grab EVERYTHING with the letters prior to the *.*, in this case, everything beginning with “Blue_Flat0” (if you’re wondering, the “0” on that file refers to the rotator being set to “0” degrees in CCDAutoPilot II). As you’ll remember we set up our Blue flats in the Flats Folder in just this way.
8) Be sure you SAVE this PROFILE, and give it a name you’ll remember (mine is “RCBlue1x1”. Again, I recommend you use my nomenclature for now (so you can trouble shoot more easily, in the event you get a program error message). Now for the big moment; hit “RUN” in the lower right corner of the dialogue window (see Figure 13 below). Before you hit run be sure the files you want to calibrate are already opened up in MIRA !!
9) After a lot of number crunching you’ll get a screen with calibration results and hopefully the message on Figure 14. click on “OK” to get rid of this screen.
10) The “CCD Calibration Results” will be seen. You can read this if you want to get a headache (Figure 15). To move to the next stage you can close some of the dialogue boxes, but save the calibrated image. Just close the “CCD Calibration Results” box ( I never save this, you may get an option to “save”) and close the “Image Calibration” box. You’ll be left with all your calibrated images, and nothing else. You need to close these dialogue boxes in order to de-clutter the screen for the next stage.
11) We are near the stage where we align the images (so they will combine without overlapping incorrectly, which would give us a lousy blurred image). BUT, before we can do this we need to 1) remove the lousy images (ruined by cloud cover, cosmic rays, poor autoguiding, etc) and 2) make sure all the images are oriented the same way (if you imaged with an equatorial and flipped the meridian WITHOUT using a program like CCDAutoPilot II, which would flip your camera 180 degrees, the western set of images will be flipped and reversed !).
So first lets deal with these 2 issues. Hopefully you closed out the extra screens, leaving your screen looking like that of Figure 16A. I put the big ARROW in, to help you with the next step (finding the “P”, see below).
12) First, we’ll orient the images. This is a real pain, which is why I use CCDAutoPilot II, which rotates my camera when I flip the meridian. If you have flipped/reversed images do the following. First get the image to a nice size, so that you can see the whole image. Do this by going to “View” and choose you magnification. With my SBIG STL-6303 I use a magnification of ¼. Next, press the button marked “P” in the near upper right (next to the letters M and G). By pressing “P” you’re telling MIRA to only rotate the image you’re looking at. Remember, we calibrated many images, as you can see in the lower near-left. Next to the red arrow. If you hit the play button you can scroll through them. Do this until you get a reversed image. THEN go to Process/Geometry/Rotate (see figure 16B, below). This will bring up the dialogue you see in Figure 16, below. First Flip, the Reverse the image (or visa versa). Don’t be surprised if the image looks VERY Bizarre (even white) on the first flip (it’s something weird which happens in some versions of MIRA, but it doesn’t effect the final result). Once you’ve flipped/reversed the image close out of the box and go to the next image. Once you have them all oriented the same way UNCHECK THE “P”. Don’t forget.
13) Now to weeding out the lousy images. Then go to the lower left and hit the play button. This moves you one image at a time. One note, you DON’T have to hit the “P” to delete one image at a time, but you DO when changing the rotation/orientation (another of those non-intuitive things). If one image looks terrible do the following to get rid of it 1) right click on the image, bringing up a dialogue box . 2) Click on the option: “Image Set” and then on “Delete Image” (see Figure 16C below). It will only delete THAT image (again, even though the “P” isn’t marked). The look of the dialogue boxes is shown in Figure 16C.
This is the stage where we align the images (so they will combine without overlapping incorrectly, which would give us a lousy blurred image).
1) OK, NOW we can align the images. We’ve calibrated them (dark/bias/flat subtracted) and we’ve rotated those that needed it and weeded out the junk. Be sure this is done now, as you CAN’T weed it out or rotate individual images once you start aligning (called “registration” in MIRA).
To begin bring up the screen in Figure 17 (below a little ways) by going to Process/Geometry/Register.
Don’t worry about all the stuff on this screen yet. What we’ll do is a 2 stage process of alignment. I’ll show all this in a moment (don’t worry). What we’ll so is first get the images close to being aligned by marking one star on each image and using the “Shift” function, Then we’ll mark 4 stars on one image, transpose the markings onto the same stars on all the images (we’ll check to make sure it worked) and then use the “Shear” function to get the images exactly aligned.
Ok, first let’s get oriented to the control panel (some new icons we need to learn to use). Most of the icons you’ll figure out by playing with the program. Here’s what you must know.
Press the “Add Points” tab (R tab) and you can mark stars for registration (i.e., alignment).
Press this “delete” tab (yellow x) to remove a single registration point (i.e., if you tried to mark a star and it ended up on the wrong spot).
Press this “delete all” tab (yellow/red x) to remove ALL registration points.
Press the “track” (diagonal x’s) tab to insert a registration point through to all the images underneath the image you just marked.
Press the “compute” (calculator) tab to have MIRA compute what it will take to align all the images (based on the registration points you placed).
Press the “Apply” (diagonal boxes) to perform the alignment process.
Press the “Preferences” (check mark) tab to tell MIRA how to do the alignment process (mainly we’re concerned about “Shift” vs. “Shear”, as we’ll see).
2) Now we’ll use Registration program to first get our images close to alignment. This will ensure that the next step will go without incident. In other words, we do this first as the automatic registration tool (the “track” tool, which we’ll use later) will not always work unless the images are at least closely aligned. Here’s what we do. Pick a bright and obvious star on the first image and mark it with the “add Points” tab (mark the star anywhere and MIRA will exactly center the registration point for you!). Do this by first clicking on the “R” and then left clicking on the star you’ve chosen. Use the forward button (or go one at a time by hitting the “tab” key on the keyboard) and mark the same star in each image. If you hit the wrong spot, fear not, just click the “delete” tab and click it over the spot with the wrong registration point, and it’ll disappear. You can remove every registration point with the “delete all” tab (if you look closely at Figure 17, you’ll see a star has been marked, with the red cross barely visible).
Once you’ve done this go to the “Preferences” tab and mark it as you see in Figure 18. The SHIFT option (under Transformation) will simply shift the stars (not rotate), but this will get them very close. It’s an extra step but you only had to mark a single big star, trust me, it’s time well spent. Bicubic is used if your Image Scale is small (e.g., less than about 1 arc-sec/pixel, which is what I use). Check conserve flux and verbose (you can read about this if you want in the help files). Then hit “OK” and the images will align. You can use the fast forward button (near the bottom left) to see if the images are at least close to aligned.
3) Now we’ll do the “accurate” alignment (registration). We do this by scrolling back to the first image. First we’ll use the yellow “delete all” button to remove our old registration marks. Then we’ll press the red “add points” button again. This time we’ll mark 4 stars in the corner and let MIRA mark penetrate through the underlying images and mark the same stars for us. Since we got things “close”, MIRA can usually do this perfectly. The process of checking MIRA’s accuracy and fixing any errors which might have occurred (rare) is a simple process.
4) First click on “R” (add points) and then left click on 4 stars somewhere near the corners (not too close…and pick stars not too bloated). MIRA will grab the center of the star, so don’t waste too much time trying to point to the exact spot of the star. If you have amazing vision, you’ll see the 4 marked stars on my image in Figure 19.
5) Now we’ll tell MIRA to put the marks on all the underlying images by pressing the “track” tab ( ). Then we’ll look to see if the stars are marked correctly. You can scroll one image at a time or hit the fast forward button and focus on the moving cross on a particular star (do this for all 4). If I find the program grabbed a wrong star I just use “delete” tab ( ) and remove that point (not the “delete all” tab, just the delete tab and click over the incorrect registration point, and watch it disappear).. Then use the “R” (add points) tab to place the red cross properly.
6) Next is an important step. We need to go BACK to the “Preferences” tab ( ). Select “Shear” under “Transformations” (see Figure 20). THIS TIME, MIRA will to whatever it takes to align the images.
6) Once this is done we hit the “compute” ( ) tab. This tells MIRA how to move the images in order to align the stars we selected. You’ll get an information dialogue box, which you can read about. I usually just close it.
7) Now for the final step. We hit the “Apply” ( ) tab. The program will perform its’ magic and an “Image Registration” dialogue bow will appear. I usually close this box. To CONFIRM that the program has aligned the images I usually hit the fast forward. The stars may brighten and dim as the images flash by, but they should not jump around. I often move the curser to a star and use it as a reference point.
1) Now that the images are aligned, we’ll be able to combine them together and get our stacked and aligned image. ALL the work we did so far was done to get us to this point. To do this we begin by opening Process/Math/Combine (see Figure 22 below). This will bring up the “Combine Images” dialogue box. Mark it as noted below (Figure 23). Mean Min/Max clipped is a good method of reducing noise. Verbose just means MIRA will display information on what it does. The help section on “Combine Images” discusses the options. Be sure to hit “Scale” and mark the “Region Stat Preference” dialogue box as shown in Figure 24 (this is discussed in the help section under “Region Stat Preferences”).
2) Close out the “Region Stat Preferences” window and hit “OK” on the “Combined Images” screen. You should see a new image which will have a header which ends with “…MMclip.fit” and a ‘verbose’ dialogue box called “Combine Images” (containing the information about the combined image).
Go to File/”Save As…” and save the image however and wherever you want. I usually have a Desktop folder with the objects name and in that folder some subfolders such as: “Calibrated and Registered Images”, “Maxim Processed”, “CCDStack Processed”, AIP Processed” and “Adobe CS Processed”. Any system which works for you is fine, since we’re about to finish with MIRA.
3) Now you’ll need to close the windows. Once you save the file I would recommend you choose “No” to any messages that say “Save changes to ….”. The reason for this is that saying “No” will insure that your raw data is untouched. THAT WAY, if you made an error OR if new processing methods become available, you’ll have your original data to work with. If you do this you’ll ALSO want to save your calibration data (i.e., everything under “All Calibration”). That’s a lot of data to save. This is why some choose to keep the changes and toss the raw data. The real professionals keep the raw data (enough said).
VI) Register the Combined Images:
1) Once you’ve repeated this process for you L, R. G. B. H-a images (or whatever you have) you’ll want to open all of them simultaneously into MIRA. First open these images up in MIRA (as with the previous steps, you need to select ALL the images you want to open and open them together). Just go to the Registration tab (Process/Geometry/Registration) and go through the alignment process again (both Shift and Shear methods). Once you’ve done this you’ll have all the images aligned. Just close out but this time be sure to SAVE the changes. The images in that folder will be ready for any processing program you want to use. You won’t have to re-align them
2) One Note: If you have 1x1 and 2x2 binned images (the most common combination), you need to add one step in the process in order to align them all together. Lets say you have L at 1x1 and R,G.B at 2x2. In that case open the 2x2 binned images one at a time and with each one go to Process/Geometry/Scale (Figure 25A and 25B below). Set X and Y to 2 (to boost it to the size of the 1x1 binned images…we’re assuming square pixels here…). Set “Conserve Flux” and “Resize Image”. Once done close out the Image Scale box (Don’t hit OK again !!!, it’ll change the size AGAIN). Then repeat this for each of your 2x2 images. Then open up the 1x1 images and do the alignment (registration) process.
That’s it. You’re done.
Addendum 1: In order to use this tutorial you will need to have a series of dark frames, bias frames and flat frames in the specified folder (which in our case will be in a folder called “Calibration STL-6303 f9 RC 15min -15deg”, which itself will be in a folder called “All Calibration”). Remember, just use my nomenclature during the first run through this program, then just change things as you wish once you have things working.
Many books review calibration basics, and this is well beyond the scope of this tutorial. As a starter I’d recommend 10 dark frames, 50 bias frames and 5 flat frames. It’s assumed you know how to obtain these (if not, Ron Wodowski’s book discusses this in detail). If you are not using CCDAutoPilot II (and I can’t imagine why anyone wouldn’t) , they can be labeled manually to match the “Source” file names discussed above. For those using CCDAutoPilot II, you’ll find an EASY way to obtain these files and have them AUTOMATICALLY named the same way I do in this tutorial.
To obtain the same file names I do for your bias/dark and flat frames follow these recommendations. I discovered them by experimentation. In CCDAutoPilot II the bias frames are obtained by shooting “lights” and setting the time to ZERO. In CCD autopilot II, under file name (in the Lights Section), label these “bias1x1_”. However, leave the name of the file blank for the dark frames and the file will default
to “-15Dark900sec1x1…”. Again, in this tutorial I have imaged everything as a 1x1 binned image, imaged for 900 seconds and -15 deg C (be sure to do so in CCDAutoPilot II if you’re saving the data automatically with this program, as I do). Also, save the bias and dark frames in the “Darks” folder, as discussed above (CCDAutoPilot II can do this as well). For what it’s worth (a lot) I’d shoot sky flats automatically with CCDAutoPilot II.
For the Flats, do not name the files and they will default to, depending on the filter, “Blue_Flat0…” Be sure the rotator is set to “0” degrees, if you want the “0” to appear in the file name (again, matching the system for the tutorial and therefore simplifying your life).
Again, once you understand how to use MIRA, you can image any way you want and change the names of any and all files. I’m simply trying to keep everything consistent while you get used to the program.
Keith B Quattrocchi MD, PhD
December 5, 2005