Part 1
In this thread I am going to briefly cover image acquisition and talk more about how I produce the picture.
My main camera is a monochrome
DMK21AU.04 manufactured by
The Imaging Source.
top of camera
rear of camera showing USB port
inside of camera showing CCD chip
It is a high frame-rate camera, capable of shooting uncompressed avi up to 60 frames per second (but I typically use 30 fps if at all possible - 60 fps can start to introduce artifacts around the edge of a planet). A typical 1 minute .avi is over a gigabyte - compare to those common DVDRip torrents that are 1 gb for a feature film!
Now you might be wondering "Isn't he talking about pictures here? Why does he have a video camera?" Planetary imaging is extremely reliant on fine details. Our atmosphere is not very forgiving in that regard. Therefore, when imaging the planets, I like to use as fast a shutter speed as possible without compressing the data. Shooting a movie is the same as shooting thousands of pictures - except I only click a record button and a stop button - not a "take snapshot" thousands of times. A fast shutter speed means there will be more potential images that are sharp from those fleeting moments when the atmosphere was still and fine detail was not distorted. The camera can only do so much though - I rely on forecasts of what the general atmospheric conditions are going to be like. The most common of these are the
Clear Sky Charts which in addition to displaying cloud cover, give a general idea of what the atmospheric transparency, stability, temperature, wind speed, humidity, and darkness (which is reliant on moonrise/set times - not city lights) are going to be. Transparency and stability almost never go hand in hand in my experience, but atmospheric transparency more impacts deep-sky objects which isn't the focus of this. Additionally, the fact that the camera is monochrome increases sensitivity: the lack of Bayer filter means that 100% of the photons from a planet reach each pixel - be they 100% red, 100% green, or 100% blue, or whatever. In a color camera the filter cuts out a portion of light per pixel, meaning that a great deal is lost.
The chip of the DMK is very small as you may have seen in the picture. Therefore, when operating at high magnifications which is what planetary imaging requires, it is more of a chore to center the planet onto the chip and keep it there. However, a bigger chip yields lesser overall magnification, and 640x480 handles almost all of what I need.
The capture program I use came with the camera - it's called
IC Capture.
Below are the "important screens" of IC Capture.
The main screen, showing all the functions. The "live preview" screen is visible. Were I imaging a planet it would be visible there.
Camera Properties window. This is where I adjust shutter speed, gain etc. I really only touch gain (camera sensitivity, same thing as ISO) and shutter speed - brightness doesn't do much and the bottom two settings are irrelevant. The other tab - "Image" only has a Gamma adjustment slider which I manipulate rarely - only if I'm seeing Newton's rings on something like Jupiter (Newton's rings are this annoying phenomena caused by low gamma where concentric rings form on the planet's disk in the final image. Kind of similar to when you photograph a tube t.v.)
This is the video recording window; has the basic functions for controlling the camera like start and stop. I can also change the video codec from here, something I never do (I've had no need to, Y800 works fine)
The histogram window. Different planets need different histogram settings. I could go on about this, but that's for a finer imaging technique thread. It's important though
I do intend to make a topic more dedicated to finer imaging practices when I have the chance. Hopefully if I see some clearer skies over Tucson I can take the scope out and create one revolving around the upcoming Mars opposition.
*********
Image Processing
The basic process of selecting the good, clear frames from the video, taking them, and transforming them into the near-final image involves one free program (although it can often encompass many more than this depending on how in-depth I would like to go, but usually I stick with this one.) The program is called
Registax 5 and is an incredibly powerful (if not occasionally buggy) piece of software.
Ideally, I wouldn't have been stupid and cleared out better videos to use for this demonstration, but I did, so all I have is a short, shitty video of Theophilus crater. I've uploaded it so if you want to follow along you can:
http://www.sendspace.com/file/fieyp3
Start Registax, and move the cursor over to the "Select" button and browse to the video. When it loads, the first frame will appear on screen and the bottom slider can be moved to browse through the movie. Alternatively, the L-R arrow keys can be used to move around.
The first Registax screen
Right by the big green square at the bottom left hand corner is the slider in question.
Now we want to fish around and find a sharp (or in this case, semi-sharp, lol) frame within the ballpark of the first frame. I liked frame 6. You may have noticed that your cursor when hovering over the video window is a box. This is to help you place what is called an "Alignment Box." You can adjust its size in the left-hand window - mine is set to 128. Choose a high-contrast feature somewhere in-frame (I chose the central peak of Theophilus) and click once. You will see a box surrounding where you clicked, with a green circle surrounding the more immediate area inside.
Frame 6 with alignment box placed
Now click "Align" and the video will play, with the box following the feature in question (this is why it is important to use a higher contrast feature, so the program can easier follow it). This should be fast because the video is short. Now if you look at the bottom slider again, you will see something like Frame (xxx) Stacksize (xxx/242) Low Quality xx.x% This is the program's quality control which you can adjust. The frame number is the lowest quality frame it has permitted. The Stacksize shows how many frames Registax has determined "quality" enough to pass over the total amount of frames, and Low Quality is the lowest quality frame as compared to your reference frame. So you can see it is important to have a high quality reference frame so the program can better determine what is good and what isn't. You can drag the slider now left or right to either cut off or allow more frames. I dragged it far-left at a cutoff of about 120 frames.
When you are done there, click Limit.
Now we are on an entirely different screen. At this point, go to the left side of the screen. If the Create a Reference frame window is not open, then click the little arrows on it to open it. Then, click the "Create" button; the frame amount is fine at 50. It should speed relatively quickly through the process and you will find yourself at the wavelets screen. These are sharpening tools, but don't get too trigger happy at this point, because that will ruin the final image. Drag the sliders right to sharpen and left to soften. I only play with sliders 2 and 3 and never drag them above 20.
the wavelets screen
You can click Do All, which will blink the image showing the image before wavelets and then the image as it is with wavelets, and then click Continue.
Back at the Optimize screen. Before we go any further, make sure that these adjustments are made:
Make sure that everything I have circled there matches your screen. That means turn Fast Optimizer off, turn Optimizer mode to V5 style, Search area 3 pixels and adjust until less than 1% improvement. Once that is done, click Optimize. This should go pretty fast as well. When it is done, click the Stack tab (directly to the right of Optimize).
Stacking helps improve detail by improving the signal-to-noise ratio. While noise is random, signal is constant which means that by stacking frames (by some weird algorithm) the noise is eliminated while the signal - in this case the finer details - becomes more contrasty. Bad frames in the stack can ruin this so in the end I cut them out.
In the Stack window, on the left, under "Stacking Options," find the check box labeled "Show Stack Graph."
There are now two sliders. The most important one is the bottom (horizontal) one, because that one is quality based. Because the video here was short to begin with, there's not much to work with. As you drag the slider left you will see the quality percentage start to go up (at the bottom of the program screen). Find a place you are satisfied with.
The vertical slider is difference based - e.g. measuring the difference from one frame to the next. You can use this to cut off any blatant peaks in the graph.
So you can see I have cut my stack size to about 35 frames, which Registax has deemed to be 90% quality thereabouts. In a more proper video this would be anywhere from 200-500 frames, depending on the capture settings used.
When you have that set, click Stack. Another quick process, now go to the Wavelets tab.
Here is where you do final sharpenings. Wavelets don't have much governing them, it's a real trial and error sort of deal. The best I can tell you is that the 1st slider is the broadest sharpen while the 6th is the finest. You can try out different combinations and click Do All to see where they take you and then move the sliders further right or back left depending on what you like.
This is what I came up with. It's a bit noisy now but I take the image into Photoshop and apply despeckle filters etc.
When you're done, click Do All and then Save Image. I personally save it as a 16bit TIFF because I take my image to Photoshop for final touches.
And that's it for basic image processing. When I next go outside, hopefully for the Mars opposition or thereabouts, I will focus more on finer imaging techniques and talk more about filters, balance in capture, finer Registax techniques along with Photoshop applications.
Habanero's how-to astronomy thread pt 2: Image Acquisition & Basic Image Processing
Re: Habanero's how-to astronomy thread pt 2: Image Acquisition & Basic Image Processi
Re: Habanero's how-to astronomy thread pt 2: Image Acquisition & Basic Image Processi
Re: Habanero's how-to astronomy thread pt 2: Image Acquisition & Basic Image Processi
Re: Habanero's how-to astronomy thread pt 2: Image Acquisition & Basic Image Processi
