M106
M106 - Click here for full resolution
Messier 106 is an intermediate spiral galaxy in the constellation Canes Venatici. It was discovered by Pierre Méchain in 1781. M106 is at a distance of about 22 to 25 million light-years away from Earth. M106 contains an active nucleus classified as a Type 2 Seyfert, and the presence of a central supermassive black hole has been demonstrated from radio-wavelength observations of the rotation of a disk of molecular gas orbiting within the inner light-year around the black hole. NGC 4217 is a possible companion galaxy of Messier 106. It is one of the largest and brightest nearby galaxies, similar in size and luminosity to the Andromeda Galaxy.
source: wikipedia
NGC/IC:
Other Names:
Object:
Constellation:
R.A.:
Dec:
Transit date:
Transit Alt:
NGC 4258
…
Galaxy
Canes Venatici
12h 18m 58s
+47º 18.2’
04 Apr
84º S
Conditions
M106 is a typical object from ‘Galaxy Season’, with best visibility in April/May. A first attempt to photograph M106 was made in April 2019, but issues with getting the mount pointed correctly made all frames captured useless. The image as presented here was captured on three consecutive nights in March 2020, under new moon conditions and with temperatures around the freezing point.
Equipment
M106 is a relatively small target, so a 1000mm refractor with a smallish sensor such as the ZWO ASI1600 is more than large enough to capture the full galaxy with room to spare. The rig was left outside for three consecutive nights. The mount model of the first session was used throughout, without further alignment during the second and third session.
Telescope
Mount
Camera
Filters
Guiding
Accessoires
Software
Takahashi TOA-130, 67FL Field Corrector, Sesto Senso
10Micron GM1000HPS, Berlebach Planet
ZWO ASI1600MM Pro, cooled to -25 ºC
1.25” mounted Astrodon LRGB filters
Unguided
Fitlet2, Linux Mint, DIY Power distribution box, Alnitak Flip-Flat, MBox
KStars/Ekos, INDI Library, Mountwizzard4, PixInsight 1.8.9-1
Imaging
In order to pick up on the faint nebulosity surrounding the galaxy core, luminance was shot with 300s frames. Even in the core of the galaxy, this did not lead to any pixel saturation. For the rest pretty standard camera settings were used, with unity gain set at 139. The image was rotated quite a bit, with the galaxy now a bit on its side. With hindsight it would have been better to give it at least a 90 degree counter-clockwise turn. This could have been corrected in processing, but then too much pixels would get lost.
Resolution
Focal length
Pixel size
Resolution
Field of View
Rotation
Image center
3031 × 1890 px (5.7 MP)
1000 mm @ f/7.7
3.76 µm
0.78 arcsec/px
49' 32.8" x 30' 53.7"
127 degrees
RA: 12º 18’ 56.008”
Dec: +47º 18’ 27.22”
Processing
At the time of publishing, the image-set was actually a four year old set of data that never really made it to a satisfactory image at the time. But with a few extra years of processing experience, and more importantly a whole new set of AI-based fantastic tools available, it was worth a second attempt.
Using the WeightedBatchPreProcessing script, all frames were calibrated using darks (50) and flats (25), registered and integrated to an L, R, G and B image. Rough edges were cropped away and each channel was background extracted using GraXpert. This is the first AI-tool used making background extraction so much easier and effective than ABE or DBE. R, G and B were combined into an RGB image using ChannelCombination.
After colour calibration of the RGB image using SPCC, the second AI-based tool was applied, BlurXTerminator (BXT). Stars were sharpened at a value of 0.35, whereas nonsellar components were sharpened using default settings. The stars were then separated using StarXTerminator (SXT), the third AI-based tool. The starless image was stretched using GeneralisedHyperbolicStretch (GHS), another great new tool. The stars were stretched using a mild ArcsinhStretch first to maintain colour, followed by a standard HistogramTransformation (HT).
The luminance image was also deconvoluted using BlurXTerminator (BXT), which really brought out some nice structure in the galaxy. Stars were extracted using SXT. Then the starless and stars-only images were stretched, using GHS and HT respectively.
Now it was time to put the luminance and RGB images together, both for the starless as well as for the stars-only image. This was done using LRGBCombination, at its default brightness and saturation settings. Stars were now put back into the starless image using the op_screen() function in Pixelmath.
The exact same data processed back in 2020 (left) and processed in 2024 (right) using the new AI-based tools available. Processing skills will probably have increased over the years as well, but these new tools are so effective and quick to use. In a fraction of the time a much better image is progress astrophotography has made over the last 4 years.
For final touches, some increase in saturation was applied using CurvesTransformation (CT). Also contrast was further enhanced using the CT tool. At this stage the image did not have a lot of noise, but the small amount of noise that was present could be effectively removed using the fourth AI-based tool, NoiseXTerminator. Only a modest amount of noise reduction was applied, using a value of 0.5. The final image is a somewhat cropped version, to enhance the presence of the galaxy a bit more.
When comparing the two images from back in 2020, vs the one processed today, the differences are striking. Without a doubt, with the proper processing skills, also the old tools and techniques could have created a beautiful image. But these new AI-based tools are making it so much easier. The new, much better image was created in only a fraction of the time that the original image took to create. There’s a couple of more old datasets that I never got to process, so would be worthwhile having a second look at those as well.
Processing workflow (click to enlarge)
This image has been published on Astrobin and received Top Pick nomination.
