This is something I'm really bad at

This image is something I'm really bad at; taking a photo with my DSLR at night. Actually, it was early this morning, like around 6:15 AM when I took the dog out. Actually, I don't really know how to use that camera for day light photography very well, either. What I saw, about 20 minutes before the image I finally took, was the moon with a really pretty circular “moon dog” around it. Visualize a bulls-eye with the moon as the center and a rainbow in the clouds as the concentric outer ring. I think the blue “thing” is what's left of the original moon dog by the time I could figure out how to work the camera. At any rate, I thought it made a cool looking picture.

Three Craters

On the night of January 19^th, I decided to start the night by imaging the moon. Lunar or planetary imaging now is done by making a movie, actually an .avi file, of the target and then processing that movie in a program like Registax. What Registax does is, among other things, “decompose” the movie into individual frames and process the individual frames. In this case, there were 500 frames, which were taken in about 18 seconds. The resulting .avi file turns out to be almost a Gigabit in size. The reason for so many frames is that a large number of frames can be used to help sharpen the final image. How, you asked? As I understand it, the idea is to use a Gaussian distribution and work back to what the image, should be. A Gaussian distribution is known as a “bell shaped curve”. If you can get a lot of points around the center of the curve, the median, and you know that the distribution of the points is Gaussian, you have a pretty good idea what number the median is, even if you don't know for sure exactly what it is. The large number of frames hopefully gives us the Gaussian distribution, which allows us to determine what the image should look like. Hopefully, you get the idea, even though that's not a really great explanation. Anyway, what we have is a slightly blurry image, which the large number of frames helps to sharpen, the Registax allows to us further sharpen the image with the magic of wavelets. If the Gaussian distribution was difficult to explain, wavelets are even harder, especially because I have at best a rough idea of how they work. Suffice it to say, the image gets sharper.

Image that has been "stacked" in Registax

Same image sharpened with wavelet functions in Registax

And what is it that we are looking at? And how do I know that's what it is?

We are looking at an area around the crater Ptolemaeus. The easiest way of determining that is to compare my image to a map of the moon. An easy one to use is the program Virtual Moon Atlas.

Here is a screen shot showing my image on the left and Virtual Moon Atlas on the right, clearly showing the same area of the moon, but VMA has kindly labeled the craters for us.

Finally, where on the moon is Ptolemaeus?

Here is a screen shot of VMA showing the full moon with Ptolemaeus labeled.

My first numbered picture (37)

Last night, January 19, I was able to get back out under the stars for a little while. There was a waxing,gibbous moon, so faint nebulae were out; too much light pollution. So I decided to aim for star clusters. It's also much quicker to photograph star clusters because the “open shutter” time is much reduced. However, I didn't realize just how much shorter the time needed to be, as we will see below. The two targets were M34 and NGC 2169, aka the 37 cluster.

First M34. M34 is an open cluster of stars that are estimated to be around 250 million years old and about 1500 light years away in the constellation of Perseus.

NGC 2169 is also know as the “37 cluster”. From the image, I think you can see why. It's about 3600 light years away in the constellation of Orion.

On the technical side, both images contain a problem know as “blooming”. It shows up as a streak of light on the brighter stars, in my case, going to the right of the star. NGC2169 has an obvious example in the bright blue star in the flat, upper “bar” of the 3. There is a dimmer star just above the streak. What's going on? The photons is a digital camera are converted to elections that are read off the CCD chip and converted to the image we see. When too many photons land on a pixel, which is where the conversion to elections occurs, they spill over onto the adjacent pixel. Think of several cups on a table with the rims touching. You pour water into the center cup, but it gets too full and spills over into the adjacent cup and starts filling it up. Although you didn't pour water into the adjacent cup, water ends up in it nevertheless. That's basically, what's happening only it's elections, not water. The cure will be to limit the exposure time to less than the 1 minute I used for these images.

For in interesting comparison, I would direct you to the blog of a friend of mine, Michael Covingtion. He has recently photographed NGC 2169, but with a much wider field of view. See:

http://www.covingtoninnovations.com/michael/blog/1601/index.html

Where's the picture?

Well, it's been a great Christmas and New Year's holiday at the observatory. That would be all except the clouds and rain. Six inches plus of rain. Finally, on January 5^th I was able to get back under the stars and see them. It was cold, however. The thermometer on the scope registered 18 degrees F at 9:30 when I came in. Never fear, though; I wasn't out the whole time. I used the Starlight Network to sit in front of the TV, nice and warm, while I let the scope, camera, and computer do it's thing; I just monitored it all. That's my idea of a good way to image. It's taken only 13 years to get to this point. My patience has been rewarded.

Well, on to the current image, which is IC342 located in the constellation of Camelopardalis. That's a mouthful. Translation, the Giraffe. This is not a constellation from ancient Greek times, but from about 1613. From Wikipedia “Camelopardalis was created by Petrus Plancius in 1613 to represent the animal Rebecca rode to marry Isaac in the Bible.[1] One year later, Jakob Bartsch featured it in his atlas. Johannes Hevelius gave it the official name of "Camelopardus" or "Camelopardalis" because he saw the constellation's many faint stars as the spots of a giraffe.[7] “ IC342 is a galaxy, very faint, of course, about 7 to 11 million light years from earth. This one pushed the scope, which has an 8 inch aperture, pretty much to the limits for my geographical location. By that, I mean that the skyglow was getting uncomfortably close to the glow from the galaxy. Allow me to demonstrate.

   

What you see above is the image of the galaxy after all the calibration frames have been applied and the luminance, red, green and blue channels have been added together, but before any level adjustments have been applied. The program I'm working in is Nebulosity, and this is the work screen shot showing the image and some information on the image. What you see in the image section are the foreground stars (those stars between us and the galaxy).

 

In the screen shot above, I've highlighted the histogram. It took me a long while to get used to working with a histogram, but now it's more or less second nature. What the histogram shows is the number of pixels of a particular value. That's rather general, so let's look at this histogram. The red part of the histogram tells me the relative number of pixels that have a 0 (zero) value on the far left of the histogram and a value of 37440 on the far right (and everything in between). A pixel of value 0 is a black pixel. As the value of a pixel increases, in becomes brighter white, so to speak. From this histogram, we can see that there is a lot of pixels nearer to 0 than 37440. As is the case with most images of the night sky, that's what we would expect; there's a lot more black than white. I can also tell that, if the image of the galaxy exists, it's in the group of pixels at the far left end. Another way of looking at it, no pun intended.....well, maybe, is that the information I want to show is in the group of pixels at the far left end. My goal, will be to use the image tools in Nebulosity to spread the small range of values at the far left end out, and into as wide of a range as I can and still have the image look reasonable. Whatever the final range of values, what I want is to see the histogram look as nearly flat, like the top of a cake, as I can get it, instead of seeing a histogram with the graph looking like it does now, with a sharp peak at the left side and little on the right.

 

What I've done is bring up the “levels” adjust dialog box. Notice that the graph there is essentially identical to the histogram in the upper right corner. What I will do is make a series of changes in the "levels" dialog and show the results. Follow along, if you want.

 

Notice that I've move the “power” slider to the left and the result in the image is that it is now lighter, like turning up the brightness.

 

Now I've closed the original "levels" dialog box, opened another and moved the black slider to the right. See how the image is now a little darker than the previous one, but, more importantly, the graph (both the histogram and the one in the "levels" dialog box) are starting to spread out. It will never be completely flat, but lets see what we can do. 

 

I've done exactly the same thing (ie, opening a "levels" dialog box, moving the power slider to the left, closing, opening a new "levels" dialog box, moving the black slider to the right) 2 more times. As you can see, the histogram is much flatter, and the galaxy is starting to emerge from the background.

 

I've done exactly the same thing (ie, opening a "levels" dialog box, moving the power slider to the left, closing, opening a new "levels" dialog box, moving the black slider to the right) 2 more times. The galaxy is clearly there, but looks rather rough. That's about all I can do with the histogram for now. What happens next will be something like... blur the image (the noise in the image is less organized than the galaxy, so blurring will help to get rid of it when I sharpen the image.), then sharpen, remove any residual background color problems (color cast), increase contrast, and so on until I'm happy with the image. After all that, the best I think I can get from this image looks like:

I hope you enjoyed the journey.