NGC2174 - Monkey Head Nebula
NGC 2174 (also known as Monkey Head Nebula) is an H II emission nebula located in the constellation Orion and is associated with the open star cluster NGC 2175. It is thought to be located about 6,400 light-years away from Earth. The nebula may have formed through hierarchical collapse. Glowing gas and dark dust do not survive well in the Monkey Head Nebula. Young stars near the center of the nebula generate stellar winds and high energy radiation that causes the nebula's material to shift into complex shapes. The nebula is primarily composed of hydrogen.
source: Wikipedia
NGC/IC:
Other Names:
Object:
Constellation:
R.A.:
Dec:
Transit date:
Transit Alt:
NGC2174 (Star cluster is NGC2175)
Monkey Head Nebula, Sh2-252
Emission Nebula
Orion
06h 09m 24s
+20º 39.5’
16 Jan
57º S
Conditions
NGC2174 is a typical winter object with peak visibility around mid January. Images were taken on February 14, 25, 26 and 28, 2023. Unfortunately most nights were hit by clouds in some way, resulting in a decent number of lost frames. On February 28 all images were lost as clouds rolled in. Moon was bright at around 90% illumination and relatively close to NGC2174.
Equipment
NGC2174 was captured using the Takahashi TOA-130 in combination with the ZWO ASI533MM camera. The Field of View of this setup was only just large enough to capture the whole nebula. But this setup was used for other targets as well during the same nights. Simultaneously the larger ZWO ASI6200MM camera was used for some wide-field images on the FSQ-106.
Telescope
Mount
Camera
Filters
Guiding
Accessoires
Software
Takahashi TOA-130, FL67 flattener, Pegasus Astro Motor Focus kit v2
10Micron GM2000HPS, EuroEMC S130 pier
ZWO ASI6200MM Pro, cooled to -15 ºC
Chroma 2” LRGB unmounted, ZWO EFW 7-position
Unguided
MacMini M1, macOS Ventura, Pegasus Ultimate Powerbox v2, Pegasus Uranus, Pegasus FlatMaster 150
KStars/Ekos 3.6.3, INDI Library 2.0.0, Mountwizzard4 3.0.0, PixInsight 1.8.9-1
Imaging
Because of the limited Field of View, centring the object was rather critical. The default GoTo from the software pointed a little bit lower in the nebula. In the annotated image below NGC2174 can be seen highlighted in the lower part of the nebula. So I made a manual adjustment and kept a frame available to align with on subsequent nights. With hindsight another solution might have been to GoTo NGC2175, the star cluster, which seems to be more central in the nebula.
Because of the clouds rolling in from time to time, several frames were lost. This led to a serious misbalance between the different filters. An attempt to get more OIII on 28 February unfortunately did not work. So it was decided to work with what was acquired.
Resolution
Focal length
Pixel size
Resolution
Field of View
Rotation
Image center
6016 ×6016 px (36.2 MP)
1000 mm @ f/7.7
3.76 µm
0.387 arcsec/px
0º 38.78’ x 0º 38.78’
159 degrees
RA: 06º 09’ 34.275”
Dec: +20º 31’ 15.29”
Processing
All frames were calibrated using Darks (50), Flats (25) and Flat-Darks (50) using the WeightedBatchPreProcessing (WBPP) script in PixInsight. In the script, the following options were selected: subframe weighting based on PSF Scale SNR, Image Registration (automatic reference frame selection), Astrometric Solution, Local Normalisation (automatic), Image Integration and Drizzle Integration (2x). So far I have not used drizzling a lot. The pixel scale of this system for example is 0.8, which would be more than enough. But the resolution of the camera is somewhat limited at 9MP and therefore I just gave it a try. The result was much more impressive than I expected. Detail and stars looked a lot better against the background of noise. So for the remainder of the processing all drizzled images were used.
The master files for each filter were quite different as far as SNR was concerned, so each was processed to a non-linear image separately. Processing included Deconvolution (automatic PSF, sharpen stars 0.2, sharpen non stellar 0.7 ) and a series of GHS stretches to bring out maximum detail in each filter. Finally the overall signal (peak of the background) was normalised between filters using a small stretch in HistogramTransformation.
Out of the individual filters two combinations were made. One image according to the Hubble palette (SHO) and the other according to a HOO palette. As there were no RGB frames to give star colour, this had to come from the narrowband filters. The HOO palette was closest to some decent star colours. Attempts to do proper color calibration on this image failed though. Instead green was removed using SCNR, which gave ok-ish star colours. The stars were extracted using StarXTerminator and kept aside for later use.
Also from the SHO image the stars were removed to create a starless image. To finetune the Hubble palette colours later on, a series of colour masks were made using the new ColorMask_mod script. A yellow, blue and green mask were created and kept aside. So far no noise reduction was applied. As next step colour noise was reduced by a simple convolution (StDev 6) of the image. This smooths out any colour noise. It also makes the image very blurry of course, but that is no problem, that will be restored by using a synthetic luminance.
As a first attempt, the noise reduction was applied to the individual filters and the detail, contrast and colour were all processed in the SHO image. However, perhaps because of the high noise levels in the OII and (to a lesser extent) the SII data, the final result was not very appealing. Some detail seem to have been lost and there was still chrominance noise visible. Therefore an alternative route was followed, where no noise reduction was applied at all to the SHO image until the end. Colour noise was removed using simple convolution. Obviously this method requires a luminance image of some sort to bring back in the detail.
Differences in color noise for different processing flows. In the left image, individual channels were noise-reduced and all image processing was done on one SHO image. In the right image, noise reduction was applied after combining the channels, by applying a strong convolution. A synthetic luminance (processed Ha-channel) was then used to bring the detail back in.
In narrowband imaging there is no natural luminance channel. But you can create what is called a synthetic luminance channel. There are different ways to do that, and I experimented with an integration of all three channels. But because they were so different in their SNR level, this did not give very good results. The Ha channel looked very good though (also the most exposure), so that was used as the basis for the synthetic luminance. The deconvolution was done a bit stronger (sharpen stars 0.25, sharpen non-stellar 0.9). Stars were removed and images was stretched multiple times using GHS. Finally noise was removed from the image using NoiseXTerminator with some strong settings (denoise 0.9, detail 0.41)
The synthetic luminance was now combined with the SHO image. Because the channels were already normalised to each other, the overall colours already looked quite nice. But still some emphasis on the yellows and blues was applied, and greens were toned down a bit. All using the color masks generated earlier. Some final touches were applied using the DarkStructureEnhance script and CurvesTransformation tool. The stars were now put back in, using the op_screen() method to give the final image.
Processing workflow (click to enlarge)
This image has been published on Astrobin.
