NGC2264 - Cone Nebula
NGC2264 - Click here for full resolution
NGC 2264 is a bright emission and reflection nebula located in the constellation Monoceros. It is also known as the Cone Nebula due to its distinctive shape. The nebula is located about 2,700 light-years away from Earth and is a site of active star formation. It is particularly interesting because it contains a young open cluster of stars known as the Christmas Tree Cluster, which is located near the center of the nebula and is responsible for illuminating it.
source: ChatGPT
NGC/IC:
Other Names:
Object:
Constellation:
R.A.:
Dec:
Transit date:
Transit Alt:
NGC2264
Cone Nebula
Emission/Reflection nebula
Monoceros
06h 40m 58s
+09º 53.7’
24 January
47º S
Conditions
NGC2264 is a true winter target with best time to observe between December and February. During this time, maximum altitudes are just over 40º in the Southern sky, creating a window of opportunity of about 6 hours per night. Images were taken on 17 and 21 January 2023 from the backyard in Groningen, The Netherlands (53.18, 6.54). Temperatures were around freezing, but the moon was absent and humidity especially on the first night was pretty low for local conditions. Especially during the second session clouds rolled in regularly, causing a high number of lost frames.
Equipment
This image is a bit of a by-product of photographing comet C/2022 E3. The comet was visible early in the morning, so a second target was selected to image in the earlier parts of the night. The comet required a fairly wide field of view, so the 0.73x reducer was used with the FSQ-106. The resulting FoV or 3.5 x 2º was way too large for the cone nebula itself. But a wide-field shot like this shows the target within a large area of Hydrogen-alpha signal with many other objects in the vicinity. This was also a nice opportunity for first light of a narrowband filter in conjunction with the QHY268c, the Optolong L-Extreme. This filter has two bandpasses, one for H-alpha and the other for OIII.
Telescope
Mount
Camera
Filters
Guiding
Accessoires
Software
Takahashi FSQ-106, 0.72x 645 Reducer, Sesto Senso 2
10Micron GM1000HPS, EuroEMC S130 pier
QHY268C, cooled to -15 ºC
2” mounted Optolong L-Extreme, Baader filter drawer
Unguided
Fitlet2, Linux Mint 20.04, Pegasus Ultimate Powerbox v2, Aurora Flatfield, MBox
KStars/Ekos 3.6.2, INDI Library 1.9.9, Mountwizzard4 2.2.9, PixInsight 1.8.9-1
Imaging
The first night had quite some technical challenges. Extremely slow interactions with the camera, non-working focusing algorithms, inability of MountWizzard4 to connect with the camera, failed dithering, etc. In the end a lot of these issues could be attributed to verbose logging turned on in KStars. With the verbose logging on for just some camera-related functions, KStars produced 4 million lines of log-output in only 5 hours. The resulting log-file was a whopping 800MB. Most issues disappeared when logging was switched back to regular again. The dithering issue was unrelated though. Since version 3.6.1, KStars works with so-called imaging trains. Since I am running unguided with the 10Micron mount, no device for guiding was selected. But when you turn on manual dither pulses, this causes an error when dithering. An error that later blocked a proper meridian flip process. Because of these issues, quite some observation time was lost though. The root cause of the dithering was only discovered on the second night, and as the system was restarted several times, in one of those restarts, the gain had defaulted to 80. Unfortunately this was only discovered the next morning. It required an extra set of darks to calibrate, the offset was not perfect and more stars were clipped in those frames than I would have liked.
Resolution
Focal length
Pixel size
Resolution
Field of View
Rotation
Image center
6199 × 4129 px (25.6 MP)
382 mm @ f/3.6
3.76 µm
2.04 arcsec/px
3º 30’ x 2º 20’
-107.1 degrees
RA: 06º 40’ 57.570”
Dec: +09º 53’ 45.68”
Processing
All frames were calibrated with Dark (50) and Flat (25) and Flat Dark (25) frames and registered using the WeightedBatchPreprocessing (WBPP) script. Since some of the frames were shot using different gain settings, two sets of Darks were used, one for each gain. In WBPP the option was selected to split the OSC image into the three different colour channels and to create integrated masters for each channel separately. Using this method allows the registration process to align the individual colour channels a bit better, to reduce optical artefacts such as chromatic aberration. With the release of BlurXterminator it is questionable if this complicated route for OSC images is still needed. BXT can do very good star corrections. The three separate masters were then combined into one RGB image.
Default stacking with WBPP resulted in dark spots (left). Manual integration with tight rejection parameters gave much better results.
The resulting image was a bit disappointing. Across the frame there were strange dark spots. Not black, but distinctly darker than their surroundings. The first thought was that this had something to do with the higher gain images, but an integration without those images gave similar spots. Blinking through the individual images indicated that such dark spots occurred on some of them. To deal with this, the integration of each channel was done manually with very tight pixel rejection values. Winsorized Sigma Clipping with sigma low of 2.0 and sigma high of 3.5 gave much better results (see image above). It is still unclear what caused these blips. It is probably a mistake in the calibration process somehow, or perhaps it has something to do with the channel separation. This is subject for further inspection later.
Next step was a background extraction procedure. But with an image that is full of nebulosity, it is very difficult to determine what is background and what is signal. So a DBE process would have been difficult. This seemed like the right occasion to test out a new tool, called GraXpert. This is stand-alone software that does one thing: background extraction. It reads the XISF files from PixInsight and can also export as such, so no image quality is lost in the process. While the software has many options for fine-tuning, the power lies in the fact that it can be used by simply clicking one button. The default settings seem to do a great job. In fact, I felt the result was slightly better than ABE in PixInsight, but again, this is difficult to judge on such a nebulous image. Overall the first impression is positive. The software is completely free of charge, so worthwhile checking out.
GraXpert is a free new software tool that does a fully automated background extraction and with good results.
Now it was time to use BlurXterminator for the deconvolution. It had a very positive effect on the stars, both in shape and colour. There was not a lot of sharpening of structure visible, even at fairly high settings. This is probably due to the fairly high noise level in the image. BXT is known to be less effective on noisy images, to not create structure that isn’t there to begin with.
I still have very little experience with processing OSC images. But Bill Blanshan has developed a series of PixelMath scripts that can be of great help in processing OSC images. They are certainly not only usable for OSC, but work well with them. So for the remainder of the processing I used several of his scripts. First is an unlinked stretching script that worked straight out of the box without any further tweaking. The next scripts are designed to work on the colours of the nebulae, and they work best with the stars removed. So StarXterminator was used to take out the stars. Then it was time to run the HOO balancing script from Bill. This does a couple of things. First it scales and normalises the channels, something that for regular SHO processing is done using LinearFit. Then it replaces the blue channel with the green channel. In an OSC camera, there are twice as many green pixels than blue pixels, so the green channel has much lower noise levels. The only colour in those channels is fromOIII and that is right in the middle between green and blue anyway. I used further the OIII boost option in the script and set it to 1.2. The result is a fairly saturated blue color for the OIII regions, that stands out nice against the background. Despite the OIII boost, there were no overexposed areas. For the green channel it blends in a bit of red. By altering one parameter in the script, the amount of red blending can be influenced. Because of this red blending in the green channel, the images take on a little bit of a Hubble SHO look. I used the default HaBlend of 0.6. The khaki/yellowish H-alpha regions were further enhanced with a few CurvesAdjustment tweaks under a yellow color mask. To enhance contrast both an CurvesTransformation S-curve was applied as well as the DarkStructureEnhance script. Now it was time to put the stars back in.
When using a dual-band filter like the L-Extreme, the stars often have a bit of a teal look to them. I’m not sure if proper colour calibration would have been possible with a dual band filter, but the teal colour could definitely be neutralised. Again from Bill’s scripts collection I used the SCNR script. This worked very effectively. Adding a bit more saturation to the stars was the only additional processing done, before they were added back into the starless image.
In the final steps of the processing some noise reduction was applied using NoiseXterminator, a tiny bit more contrast was applied, and the edges were straightened out a bit using a mild crop.
Processing workflow (click to enlarge)
This image has been published on Astrobin.
