Camera tutorial

Dominik Gronkiewicz

The main concept of planetary cameras is to deliver a high-rate sequence of frames at the cost of bit depth of the A/D unit and S/N ratio of a single frame. Such cameras are used for imaging bright objects of our Solar System. Instead of taking a single exposure, each photo is obtained by registering a few hundreds of frames. The best quality frames are picked, aligned and stacked. The process of stacking turns random atmospheric distortions of individual frames into uniformly blurred image. If needed, this may be corrected using standard deconvoluntion algorithms with gaussian kernel.

NOTE. You probably want to read read our general guidelines for solar photography.

Capture

NOTE. I use free FireCapture software for image acquisition and prefer it over the software delivered by my camera manufacturer. Although this tutorial is based on FireCapture, any software that is able generate observation logs will be good for our purpose. Logs are needed to determine the precise time of observations.

NOTE. In case of any trouble, go straight to FireCapture help center.

0. Make sure your laptop battery is charged and that you have a few tenths of gigabytes of free memory. Synchronize your laptop clock using synchronization software.

0b. If it's your first time using FireCapture or you never paid much attention to file naming, that could be a good moment. You can change the default file naming in Settings tab. I recommend the following configuration:

0c. Fill in the information about you, your location and the telescope used. FireCapture will automatically place it in the log files. That is going to make data handling much easier for future users.

1. Set up your telescope. If your observatory is mobile, align your equatorial mount as good as possible. You may use a compass, but remember to take magnetic declination into account.

2. Prepare your telescope. Put an objective filter if you're going to do white-light pictures. UV/IR cut filter is highly advisable if your camera doesn't have one.

3. Point your telescope at the Sun and install the camera. Plug it to the computer and run FireCapture.

4. If your camera doesn't capture the entire solar disk, point it at one of the active regions visible on the Sun. This is where we expect the flares to occur. We usually suggest what's hot on the Sun in the news on our website. Try to keep a sunspot in the center of your image. If no active regions are present, you may select any other feature.

5, Adjust the sharpness and lock it with a screw. If you use a tension-tuned H-alpha filter (etalon based, for example Coronado PST), turn it until your active region is inside the dark „sweet spot”. Inside this dark area, your image is tuned precisely to the center of H-alpha line. You want to keep your target in this dark region throughout the observations. Users of better quality etalon filters or thermally stabilized filters don't need to worry about the „sweet spot” effect.

6. Adjust the exposure so that the histogram is filled in about 30-40% for Ha and 70% for white light. Keep the exposure time short and make sure that your camera is working at a maximum framerate. Set the gamma coefficient to 100 (absolutely required!) and reset brightness, contrast and other adjustments. We want our data to be as raw and linear as possible.

7. Choose the amount of frames to capture. I usually set this parameter to about 300 frames which is fairly enough if you don't use crazy focal lengths and no harsh processing will be done. More frames means better signal but also more disk space, which may be an issue during a few hours of observations.

8. At this point you might want to think about the guiding. There's little chance that your mount can guide the sunspot at high magnification for a few hours. If you plan to do the guiding manually with a hand controller, the „Reticle” option will display a cross that I find extremely helpful to keep the image on the spot (see image above). Once you feel the directions, guiding for extended periods of time may be a bit boring, but definetely not very exhausting.

If you have a mount that is able to communicate with computer via ASCOM platform, good news – FireCapture can do the guiding for you! First, activate the „Align Box” feature it in the right panel and go to „Auto Align” section in the Settings window. „Surface” is the mode you're going to use during solar work. Play with „Threshold” as well as other sliders, to make the wobbly red rectangle surround one of the visible features (for example, a sunspot umbra – see picture above). This is the reference point for tracking the mount.

Now, go to „Telescope” tab of „Settings” window and establish a connection with your mount (you know the best how to do that). We're almost done! Once you check the „Autoguide” option in the main program window, FireCapture will attempt to keep the red rectangle fixed in its position. I used the word „attempt”, because most probably you will have to play with autoguiding options for a while to make the guiding accurate yet prevent FireCapture from rocking your telescope back and forth like crazy. However, this has to be done only once. FireCapture will remember the parameters until the next imaging session.

9. If everything is double-checked and ready to run, then open the old good „Settings” window and go to the „AutoRun” tab. Set „Number of runs” to desired number of captures (you can break the sequence earlier if you need to). Now, how to select a reasonable delay between runs? The desired imaging cadence (which is capture time PLUS delay) is 10 to 20 seconds. So, for example, if you want to get imaging cadence equal to 20 seconds, each capture being 5 seconds long, then the „delay between runs” must be equal to 20-5=15 seconds. Try to capture as long sequence of movies as possible (up to a few hours).

If you're REALLY ready, then click the „Start” button, and the machine will run...

10. Depending on how well you aligned the mount, you'll probably need to correct the telescope pointing every few minutes. Guiding Sun in the center of the view is important in all telescopes to keep aberrations low and to make further processing less problematic. However, as we said, accurate alignment has a crucial importance in H-alpha telescopes, where the „sweet spot” of good filter transmission barely covers entire solar disk.

Don't change exposure parameters once you start taking photos. The only exception is when you notice an image saturation during a strong flare – you may decrease the exposure time for the maximum phase. However, remember that each setting change will require a separate dark frame, so use as few different times as possible.

11. After you finish the session, do not change exposure parameters in your camera. Point your scope away from the Sun. Cover up your telescope and record a „dark movie” which will contain nothing but camera noises. We're going to prepare a „dark frame” from this movie and subtract it from our image. We describe calibration frames in introduction to photography section.

12. Look at the nasty dirt on the CCD chip of your camera. How to remove it?

If you're imaging in white-light, point the telescope at the clear blue sky in the zenith (without the solar filter). Adjust the shutter speed so that the image is neither too dark nor oversaturated (you can easily check it by looking at the histogram – should be in the middle of the scale). Once you're done, take a „flat movie”. We need such view of a uniformly illuminated surface (sky) to map all imperfections of your telescope's optics as well as dusts on the camera's sensor which are clearly visible as dark blurred circles. Such a frame is called a flat field. We're going to use it to remove even the smallest dust from the image!

If you use a Ha telescope, you probably cannot remove the filter. However, if the field of view is narrow enough, you can (and should) take the flat field! Simply find a relatively empty area near the center of the solar disk (limb areas are highly not recommended) and blur the image slightly. Voila! Record a „flat movie” and proceed to the next step.

13. Once again, without changing exposure parameters, close the telescope lid and take another „dark movie” to collect the noises for the flat fields you have just taken. Usually, exposure time for my flat fields are much longer than for imaging the Sun, so separate dark frame is necessary.

Now go home and make yourself a cup of cafe latte to make the processing more pleasurable.

Processing

NOTE. We're going to use AutoStakkert for processing (click here for download). I recomment the latest beta version. You're free to use any other software that you like. However, AutoStakkert is very convenient when it comes to processing large datasets. Also, AutoStakkert is able to open SER files as well as AVI files compressed with UT-Video lossless codec.

AutoStakkert is quite a friendly software. It's extremely simple and ergonomic. It is also fast and its algorithms are considered among the best.

The graphical user interface is limited to only two windows. The main window is used for most operation. It has many options and switches. The layout is very intuitive and options are arranged in proper order.

In the second window, the video currently being processed is displayed. You can jump between frames using the slider on the top.

Here is the list of steps of our processing:

1. Prepare the master dark frame

2. Prepare the dark frame for the flat field

3. Prepare the flat field

4. Load the movie

5. Take a look at the quality

6. Throw the anchor!

7. Have a coffee

8. Send us the results

Step 1. Prepare the dark frame

Click Open and load a “dark movie”. As you can see in the picture, I tend to give files proper names, and I think that's a very good habit.

What we want to do is to compute a mean image of noises (the proper dark frame). This can be done easily with Create Master Frame command.

AS!2 is going to display a window, asking you to specify a name for destination file. I'll name it dark.tif. Pick whatever you like.

There's not much noise here, however, the constant offset level will be substracted.

Step 2. Prepare the dark flat

The procedure of preparing the dark frame for flat field is basically identical. Just remember to pick the correct file! Again, good file naming will help.

The result is a bit more noisy because I had to apply longer exposure time for a flat field.

Step 3. Prepare the flat field

We are almost done with calibration. Load your flatfield. You're probably going to see lots of dirt on the detector.

Remember that before we stack this into one frame, we must substract the dark frame. Load the dark frame created in Step 2 using Load Master Dark command.

The presence of dark frame will be indicated on the central panel.

Now just create a master flat field with Create Master Frame command. Eww, look at that dirt on CCD!

Step 4. Load the movie

Once again, click Open and select all AVIs in the sequence. This is a great feature of AutoStakkert!2 – it allows you to process many files at once.

Beautiful. Now let's load our brand new dark frame and flat field by choosing appropriate options in Image Calibration menu.

Once again, correct loading of calibration frames will result in notification being displayed.

Also, as you load the flat field, you should observe the dirt on the image to disappear:

Now, some serious stuff begins. AS!2 offers two work modes, dedicated for planetary processing and extended objects (like Sun or Moon). Make sure that Surface switch is checked. You also want to Crop the stack size because data from image edges is useless anyway. Noise robust is an important parameter that tells AS!2 about the scale of visible details. Low values (1,2) correspond to very sharp images at excellent seeing conditions and allow AS!2 to analyze single pixels. The noisier and less stable the movie is, the higher the value should be, forcing AS!2 to judge the quality according to larger details. Value 3 works best for me. The last switch defines the way of quality management. It should be set to Local, unless there are clouds passing though the image (the author recommends to use Global option in such case).

NOTE. I'm not going to get into too much detail about each feature of this software. You might want to check an appropriate section on AS!2 official site.

Now, go to the second window. You must select a characteristic feature (such a sunspot) that AS!2 will use to track the image as it moves during the capture. Click that feature with ctrl button pressed, just as the instruction says.

Step 5. Have a look at the quality

Now push the Analyse button. AS!2 will analyse the quality of the frames and sort them from the best one to the worst one. You can swing the slider from left to right to see that the above is true. We must reject some part of the worst frames, the only matter is how many of them. It's also useful to look at the graph in the middle that shows a quality estimator. In the example below, there's a significant drop in quality past 25% of the frames.

In the top right part of the window, you can select the output format (please pick TIF, as FIT seem not to fully work yet). The pink/green fields are one of the most important elements of the program. You may specify the number/percentage of frames that will be used for stacking. Reasonable values are between 10% up to 80%, depending on seeing conditions. Limiting the number of frames in bad seeing will improve the image. You can type more than one value and AS!2 will generate frame for each setting, which is excellent for testing.

Seeing conditions during taking example images were absolutely awful, this is why I decided to use only 25% of the frames for processing.

You also have to specify the size of a reference frame, which is a frame stacked from a few best images and used for alignment. Using a reference frame guarantees better geometry preservation (which sometimes is a problem during multi-point stacking). I recommend using 50-100 frames for a reference frame (more will only blur the image).

Step 6. Throw the anchor!

Actually, not only one anchor, but tenths of them. Try clicking on the image. Green squares will appear. You can change their dimensions using mouse scroll or buttons on the left side. These squares are so-called alignment anchors and they have a very important role in processing. As you have probably noticed, atmospheric seeing causes different part of images to fluctuate. By arranging many anchors on entire image, we can track these changes and un-warp frames, partially cancelling the effects of atmospheric seeing.

When you're having a walk in a historic city center, you use characteristic points for navigation. On the other hand, one can get easily lost on the sea, desert or forest, where there now orientation points available. Our case is similar: place the alignment anchors over characteristic points, such as sunspot, prominence, plage or granulation (if visible). The fainter and more spread the detail is, the bigger anchor is required.

NOTE. This point explains why not using flat fields is a major mistakes made by solar photographers. Any dust left on the image will probably confuse the aligning algorithm and cause some part of the resulting image to be aligned improperly.

Manual placement of anchors is good idea in Moon and planetary processing. From my experience in solar processing, if granulation, spicules or other surface details are visible, it is enough to spread anchors uniformly on the image area. (The lack of subtle disk details on the image may indicate poor collimation, optics quality or poor seeing conditions.) Set an anchor size to a reasonable value (I usually use values of about 60 pixels) and click Place APs in grid.

Entire image should be filled with an array of anchors:

Step 7. Have a coffee

The last options configure the behavior of AS!2 after processing, particularly – file naming.

If you want to, you can switch on the Sharpened images option, which will generate sharpened copy of each picture. However, these images are for preview purposes and we ask you to send us only raw stacks.

I recommend to tick Save in folders as well as Simple filename for RAW options. The former will close processing results in separate folders, while the latter prevents AS!2 from heavy modifications of file names. If you configured FireCapture to name AVI or SER files nicely, then all file naming is done automatically for you.

The last option that you might consider is using HQ Refine together with Drizzle 1.5X. This is algorithm, originally designed for Hubble Space Telescope that allows to re-sample images and extract more information (unlike ordinary resizing that doesn't actually improve the data). Drizzling may be applied to excellent data, taken at very good seeing conditions and with instrument of excellent sharpness (better than pixel size). Below I compare two raw stacks from one of my best imaging sessions. Note that despite no sharpening granulation is visible well.

If you're ready, click the Stack button and take a break. Depending on number, length and resolution of the movies, as well as abilities of your computer, it may take a while to finish the processing. Luckily, AutoStakkert!2 is fast and behaves very well during work, which means that you can cancel the processing at any time without hanging up the application.

Step 8. Send us the data

AS!2 will place result stacks in subfolders, one for each stack size (in my example, named AS_p25). This is what we would like to receive from you!

Apart from these files, we need some additional information about the acquisition itself. Luckily, it is contained in text files, generated automatically by FireCapture. (This is why this software is good for our purpose.)

Here is an excerpt from one of the log files:

Camera=DMK 21AU618.AS
Filter=540nm
Profile=Sun
Filename=Sun_20141104_131902_540nm.avi
Date=20141104
Start=131902.765
Mid=131904.768
End=131906.772
Duration=4.007s
LT=UT

As you can see, these files contain a lot of precise temporal information that is needed for scientific analysis and comparison with other instruments.

Text files are very light, so just put them together with TIF stacks to one archive and submit to F-Chroma.