Monday, January 14, 2019

#175 - Saturday, January 5, 2019 - Winter Fun

I definitely put "2018" in the title of this post and had to change it to 2019.  Oh January!

Astro-buddies do more than just show off views in each other's scopes and try to find each other by voice alone in the dark -- we also sometimes hang out during the day and do non-astronomy stuff!  A group of about 15 of us went to the same park that our club's observatory is located in and went for a hike along the scenic trails.  It was muddy, but it was a sunny day with temperatures closing in on 50 degrees, so it was beautiful!

Me crossing an overflowing creek with the help of my trekking pole!

Because of all the rain we'd had over the prior couple of days, the creeks and river flowing through the area were pretty high and rather muddy, which made for some extra adventure!  After the hike, we all went and had lunch at a Caribbean-themed local restaurant, which was delicious.

Lunch ran long because of the large group, so I dashed home to grab my gear and return to the park for member's night.  The skies were very clear, but the forecasts were mixed on whether there would be thin, high clouds, so I couldn't decide whether I wanted to bring my telescope out or not.  Ordinarily, especially during the winter, I like to use my club's memorial telescope in its own dome on the observatory property, since all I have to do is plug in my camera and get on target.  But it's open for everyone to use for member's night, so I had to decide whether to bring my own telescope rig, or just do widefield imaging on my Vixen Polarie.  By the time I got home, I'd made up my mind - I would be crazy to not  bring my scope out with that good of a possibility of it being clear! How lazy have I become??  Especially since my Celestron AVX + Borg combination is so light and easy to set up.  Now, that did mean I had to bring a lot of stuff: the mount box, mount tripod, Borg case, ZWO and guide camera case, Nikon and Vixen Polarie camera case, camera tripod, tool box (giant tackle box that has everything from tools to adapters to velcro to batteries), accessory box (bin with power strip, dew heaters, extension cord, AC-to-DC power converters, etc), folding table, and chair.  But it would be worth it if the skies stayed good!

I didn't get out to the park until nearly 6 PM, and member's night started at 5:30.  Since darkness was coming fast, I hurried to set up both the AVX and the Vixen Polarie.  I set up the Polarie first, and worked on the AVX while test frames were coming in so I could hone the polar alignment.  My plan for the Polarie with my Nikon D5300 riding atop was to image the lower half of the Orion constellation, including the Great Nebula of Orion, Flame, and Horsehead Nebulae, and any other nebulosity I could tease out of the background in this very dusty and colorful region.  After messing with the polar alignment until the streaking in the stars was minimized, I tried for 3-minute subframes, but that was still just a big too much for it even at 100mm of focal length, and so was 2 minutes 30 seconds.  I wanted some detail in the nebulae, so I needed very little streaking.  I settled on 2 minutes, set the ISO to 1600, and let it go.  Orion was still too low, but then I could focus on the AVX and I would just delete frames later.

Single 2-minute subframe on the lower half of the Orion constellation

I've been meaning to apply the same backlash fix to the RA axis as I did the dec axis, but haven't gotten to it yet -- I've been suuuuper busy lately.  In addition, the dec axis was weirdly tight (I still don't think I have the clutch knob in exactly the right position), so I couldn't really balance it.  After getting it built, the first thing I did was polar align with SharpCap, but that proved difficult because the backlash in RA meant that the field kept moving up even as I was waiting for frames to come in between mount adjustments.  As a result, my gotos weren't very good at all.  I finally bought a dual-finderscope mount so I could have my red dot finder and my guide scope loaded at the same time, which would have helped the alignment process go smoother if I was actually able to boresight the red dot finder.  The bracket holding it is somewhat tall, and I can't get it to adjust low enough to be looking at the same spot in the sky that the scope is.  I'll need to find a shorter dual bracket...or a more-widely-adjustable lightweight finder.

After aligning, I used Precise Goto to put a target that has been on my list for a while now, but keeps not happening for one reason or another -- the Cone Nebula.  But since my gotos weren't good, I wasn't totally confident it really did put it in the middle, and I haven't re-set-up plate solving on my replacement tablet yet.  Besides, I didn't want it in the middle -- there's some more cool-shaped nebulosity and a cluster just above it, which I wanted as well.  Conveniently, there's a 5th-magnitude star right in the center of where I'd want the frame to be, 15 Monocerotis, which I was pretty sure I could identify in the frame, so I centered it.  I started taking 3-minute luminance frames on my ZWO ASI1600MM Pro, but I couldn't see it in the raw FITS files.  However, it's not a terribly bright thing, so I decided to cross my fingers that it would come out in processing and pressed forward with imaging, even though there was a chance I wasn't even on it.  Guiding was looking good, however.  With everything rolling, I went inside to warm up my freezing fingers and toes and finally have some conversation with my sky people.

30 luminance frames later, I went outside and carefully flipped to the red filter.  I also grabbed a pair of binoculars so I could take a look at the Great Nebula of Orion, M42, since the sky was pretty darn clear.  I was so glad I had brought out the scope!  The transparency was good enough that many of us could make out the wintertime Milky Way up high, and another club member was able to spot M33 in binoculars.  People looking through dobs and binoculars were ooh-ing and ah-ing at M42.  

I came back out later on to switch to the green filter and saw that the laptop I was lending my minion Miqaela, whose laptop is out of commission at the moment, had fallen off the table!  It was in a tub, so it wasn't on the muddy ground at least, but I had no idea how it'd gotten there.  Then I saw that my chair had fallen onto my mount tripod as well!  It was a bit breezy, so there must have been a gust of wind that knocked a few things over.  Luckily, my camera tripod with the Polarie didn't appear to have moved at all (I hung the AC-to-DC adapter for my camera off of the tripod hook, which helped a bit).  The alignment was off a bit, so I re-added the alignment stars, which happened to be on the same side of the meridian that the Cone Nebula had just crossed into, so I didn't bother with the calibration stars on the east side of the meridian. I did a Precise Goto, but I don't think it got me quite to the same spot.  But the sky quality was declining anyway, and it was getting late, so I decided just not to deal with it because that would mean re-polar-aligning and re-aligning, which I didn't feel like doing in the cold.

Around 12:30 AM, the clouds started rolling in, so it was time to pack up.  We got out of there by 1:30 AM, and I was in bed by 2:20.  Luckily it was a Saturday night!  I'll be curious to see if the Cone Nebula is in my luminance images.  If not, I at least have my Orion constellation widefields to enjoy!

[Update January 13, 2019]


I started with processing the Orion constellation ones because I'm pretty excited about them.  I won't do an exhaustive step-by-step here because I have that in other posts, but I'll outline the workflow in PixInsight 1.8.6 and talk about anything special or new.

First, I created master dark and bias frames (slowly working my way through my library creating these as needed), and the calibrated the light frames with these.  Then I debayered the light frames.

Subframe Selector

Since I had 127 frames after deleting the ones that were too cloudy or too low in the sky, this dataset was a good candidate for me to weed out even more non-ideal frames.  A handy way of doing this is the SubframeSelector script.  One of the parameters it requires is your camera's e/ADU, or electrons per analog-to-digital unit.  When your camera is reading out the image on the sensor, an analog signal of voltage (which was originally electrons collecting on a capacitor) is converted to digitized units.  This conversion depends on the gain setting of the camera.  While astro CCD cameras usually tell you what the e/ADU is for a few different gain settings, this is not the case for DSLRs, and I have long struggled with just having to make a guess of this value.  However, upon further inspection of the Light Vortex Astronomy tutorials, I found a page that talks about a PixInsight tool for calculating it -- because of course there's a PixInsight tool for this!  It's called BasicCCDParameters. The link for the Light Vortex tutorial is here.

I went and located flats, biases, and darks all taken at ISO-1600.  I got biases and darks at the same temperature - 58 degrees - but it's been a while since I've done flats, so I hunted around and found some at 48 degrees.  Hopefully that's close enough.  I also needed to find out what the maximum ADU was, so I needed an image that was saturated.  I remembered that some pictures I took on a road trip through Glacier National Park in 2017 had come out saturated, so I opened up one of them in PixInsight to check is max value.  To do that, I used HistogramTransformation, changed it to 16-bit mode, and looked for the saturation peak.  It was indeed at 16,383, which is what the BasicCCDParameters script detected initially anyway.  I loaded them all into PixInsight, checked the parameters, and hit Report.

It's so interesting to finally see these numbers!  The gain e/ADU that I was looking for is 0.115 e/ADU.  Also worth noting are the read noise of 1.465 electrons, dark current of 0.002 electrons per second, and full well capacity of 1,889 electrons.  That well capacity is quite low (which is not ideal), but the dark current is very low (which is ideal!), like way more than I would have thought!  Same with the readout noise.  My ZWO boasts a read noise of 1.2 electrons, so about on par.  CMOS for the win!

With that value in hand, back now to SubframeSelector and followed the Light Vortex tutorial about it.  After my computer crunched away at analyzing the frames, I first looked at the plots to see how things looked.  SNRWeight was interesting- you can see it increasing throughout the course of the night.  This is a direct result of the haze that slowly drifted in.

After looking at the plots, I made some cuts, using the following expression in the Approval box:
FWHM < 2.9 && Eccentricity < 0.89 && SNRWeight <=3.3
This cut out 23 of my 127 frames.  A reasonable number.

Then I used the expression from the tutorial to calculate the weights.
(15*(1-(FWHM-FWHMMinimum)/(FWHMMaximum-FWHMMinimum)) + 15*(1-(Eccentricity-EccentricityMinimum)/(EccentricityMaximum-EccentricityMinimum)) + 20*(SNRWeight-SNRWeightMinimum)/(SNRWeightMaximum-SNRWeightMinimum))+50

Finally, I went back to the table of values that are shown in the plot and sort by weight so I can find which one is the best subframe.  It's DSC_0172, so I opened it first (the original file using the Windows photo viewer, since you can't open files in PixInsight while you have a script open) to check for airplane or satellite trails.  It was clear, so that is the one I used as the registration reference frame. 

Making it pretty

Once registration and stacking (using the weights as a keyword) were complete, then it was time for DynamicBackgroudExtraction.  Selecting the sample points for this process was tough -- first, I had to decrease the sample size to 10 pixels because there were so many stars in this image, and I had to adjust the location of almost every one of the samples to not be on a star.  Second, this is a very nebulous region, and I'm trying to avoid the dim background nebulosity I can't see so that I can bring it out later.  I had to delete nearly all of the points in the lower right quadrant because I could see nebulosity in them. It shows up as colorful pixels between what are obviously stars in the sample display box of the DBE process.

The result was very exciting!

Look at all that dark nebula!  And that's just a screen stretch!  Get excited!!

The rest of the process was what has now become mostly my workflow:
- Color calibration with PhotometricColorCalibration
- Denoised with MultiscaleLinearTransform, with a lightness mask (stretched)
- Created star model with DynamicPSF for use with Deconvolution
- Created range mask and star mask from already-stretched lightness mask, subtracted two
- Used combo mask for applying deconvolution, 30 iterations
- Stretched with HistogramTransformation
- Contrast enhancement with HDRMultiscaleTransform, with range mask
- CurvesTransformation to touch up curves
- ColorSaturation - boosted reds
- Ran the DarkStructureEnhance script, with lightness mask
- Additional denoising with ACDNR

Aaaaaand here it is!!

Date: 5 January 2019
Object: Orion Widefield
Attempt: 3
Camera: Nikon D5300
Telescope: 55-200mm lens @ 85mm, f/5.6
Accessories: N/A
Mount: Vixen Polarie
Guide scope: N/A
Guide camera: N/A
Subframes: 104x120s (3h28m)
Gain/ISO: ISO-1600
Stacking program: PixInsight 1.8.6
Stacking method (lights): Average, linear fit clipping
Post-Processing program: PixInsight 1.8.6
Darks: 100
Biases: 20
Flats: 0
Temperature: 41F (approx - didn't have thermometer)

I'm very excited about this result!  The core of M42, the Great Nebula of Orion, is blown out as expected, but between having excellent focus, good tracking, and using SubframeSelector, I got some very nice detail for how large my pix scale was (due to the short focal length).  In addition, I picked up some dusty regions, including the elusive Witch Head Nebula in the upper right, which is truly incredible considering the level of light pollution I have to deal with.  Also featured are the Flame Nebula and the Horsehead Nebula at the bottom of Orion's Belt, the Running Man Nebula right beside M42, and the small nebula M78 near the left edge of the image.  See the AstroBin link above for a full list!

I haven't had a chance to see what I got out of the Cone Nebula yet, but hopefully soon I can process the luminance and see if I was even on target!

[Update January 15, 2019]

I wanted to take a look at the luminance data, but in case it was good, I went through the whole process of creating the master dark and bias, calibrating the lights, using SubframeSelector do eliminate less-than-ideal frames (kept 20/25), registering, and stacking.  Aaaaaaand there's nothing there. :O

Just a dust spot and a bunch of stars.  Rawr!  Next time...

I wanted to find out where I was really pointing, so I uploaded the stretched jpeg to  After it finished solving, I clicked the button to view it in World Wide Telescope so I could see where it was on the sky.

No comments:

Post a Comment