Very Amateur Astronomy

Well that was a long wait for a clear night, what an unproductive three months for astrophotography but at least the front garden looks nice in it’s new sandstone paving! Sunday night finally delivered the required combination of clear skies, less than half a moon and a following morning I didn’t need to get up to early so I set up and finally got to have a go at a widefield image of the Cocoon nebula and the dark nebula which seems to stream away from it.

The Cocoon itself is a combination or emission and reflection nebulosity, the dark nebula that stretchs out from it is a cloud of dense dust blocking out the light from the stars behind it.

Imaging Equipment

• William Optics ZS66SD
• William Optics MkII Field Flattener
• Canon EOS 1000D (modified)
• Astronomic CLS CCD Filter

Guiding and Mount

• Orion Optics 250mm F4.7 Reflector
• Opticstar PL-131C Camera
• IR/UV Cut Filter
• Celestron CGEM Mount
• PHD Guiding

Exposures

• 44 x 5 minutes
• 15 x bias
• 15 x dark
• 15 x flat

The images were calibrated, aligned, stacked and processed in Pixinsight.

With the processing of this image I took the opportunity to try some new processing techniques within Pixinsight with the aim of improving noise reduction and getting better colour from the image. You can see some of my other attempts from the NGC and Other Catalogs gallery page. The first image is from my standard technique which ended up with a poor colour balance, the second was better in that respect but suffered from noise issues especially in the regions of the dark nebula.

In the version shown on this page I used the following techniques to try and get a better result, the links will take you to the Pixinsight forum post which discusses them :

Wavelet-based noise reduction with the ATrousWaveletTransform tool

I have always used the ACDNR tool in Pixinsight before to manage noise reduction post histogram stretch / sharpening when the image is in a non-linear state. Noise reduction from this alternative newer method however can be applied to a linear image and works as many tools do in Pixinsight by working different structure size layers. This means your noise reduction can be tailored to match the level of noise you have, hopefully that way avoiding blotchy backgrounds where larger areas of noise have been let through as structure.

I was very happy with the result once I had spent some time on trial and error to achieve the best result for my image.

RGB Re-Combination using Linear Fit

Technically I am pretty sure this is not for us single shot colour DSLR imagers but I cannot complain at the result. The technique here is to do a rough stretch of your Luminance data (I used an extracted lightness image) that looks about right, then match that stretch on your RGB images before applying Linear Fit tool to the RGB to match them against a reference image. I still had to use the SCNR tool to remove excess green from the final image but I have to call myself happy with the end result. Colour seems easier to manage, and the red of the Cocoon is closer to red than my normal orange. It is not scientific and probably not right for my data but as an experiment into alternative processing I like the result.

Hopefully the wait between new images can be a bit shorted this time!

We are well into galaxy season now and the weather has been kind for most of the week, so queue some new images. First off I wanted to take a better look at NGC2903, my previous image had been severely hampered by poor seeing. The raw unprocessed (resized) images below show the difference, first is an image from the 28th of March :

And the second one taken last night :

The problem with the first image was a thin haze in the air which was catching the light pollution, it does nothing for your imaging, washing away fainter details.

Anyway, with several clear night to choose from I started over from scratch on NGC2903, capturing 42 new images through evenings of good seeing throughout the week. I think the result shows clear improvement.

NGC2903 Barred Spiral Galaxy in Leo

• Pixel scale:1.83 arcsec/pixel
• Field size :30.55 x 21.20 arcminutes
• Field contains:
  • NGC 2903

Exposures

• 42 x 5 minutes
• 21 x bias
• 21 x dark
• 21 x flat

With a fully clear night last night and no plans for today I calculated that I could capture another hour on NGC2903 and leave myself four hours for a second target. After consulting my new copy of the excellent ‘100 Best Astrophotography Targets‘ by Ruben Kier I decided that NGC4565 looked like a suitable target, both due to it’s position in the sky which would allow the full four hours and for it’s apparent size. Guiding went smoothly throughout the night as did the meridian flip.

NGC4565 The Needle Galaxy in Coma Berenices

• Pixel scale:1.71 arcsec/pixel
• Field size :29.19 x 20.87 arcminutes
• Field contains:
  • NGC 4562
  • NGC 4565

Exposures

• 48 x 5 minutes
• 21 x bias
• 21 x dark
• 21 x flat

I have spent more time in Pixinsight on both images trying to get better colour from them, I think the results are good on that front. Normally I use an extracted lightness mask when performing the saturation stretch, but found in this case that it was actually masking off the fainter parts of NGC2903, improvements to that process allowed me to bring out the blue of the fainter spiral arms. A similar technique was used for NGC4565, there may be a better image still hiding in the data but for now I am happy with the results.

I was actually pretty happy with my recent images, I thought they came out pretty well. However looking at ngc2403 and ngc4244 I started to wonder where the colour was, the modified camera should be picking the regions of nebulosity much better than it appeared to be, and the images were rather more blue than I would expect them to be.

I reopened each of the stacked but unprocessed images in Pixinsight to have another go and see if I could work out what was going wrong. Previously I had improved my colour correction by ensuring that I did not normalize the image during the dynamic background extraction process, but I had been neglecting to make use of the colour calibration tool after this, trusting instead to the colour balance of the modified camera combined with the Astronomik CLS-CCD filter, plus some manual tweaks later on in the process.

This time round I went back to Harry’s excellent Pixinsight tutorials and combined the colour calibration lesson into my own process. It really does help to go back to fundamentals some times, a refresher for the brain to remind yourself why you are working in a specific way. The results can be seen below, I think compared to the original images posted here and elsewhere these versions show significant improvement. Most notably in the galaxies where the ‘nearby galaxy’ method of colour correction has done wonders.

NGC4244 – Significant improvement, obvious colouration in the centre of the galaxy, better star colours and better background.

NGC2403 – Shows greatly improved colour, areas of nebulosity are standing out very nicely now, though the image lacks some sharpness I think.

NGC2174 – Colour calibration was completed via the aggregated stars method, not such a massive change as seen in the images above but it is certainly improved. Greater contrast and range is visible.

An exercise in using the least scientific or structured method possible to pick a target, run through a bunch of likely galaxies in the same area of the sky and pick the one that comes with a nice bright guide star already lined up! I’m not proud but it does save me time and frustration.

We’re very much heading into galaxy season now, where the chance to image nebulas is replaced with the chance to image faint and often very small galaxies instead. At least small for my field of view, it does present a problem. Once I have imaged all the targets which fit nicely in my field of view, be they the very few single galaxies or the nice groups I am left chasing very tough, small targets which may give me a nice challenge but don’t give a particularly impressive photo at the end.

The other options of course is to go back over the better targets with improved equipment and more knowledge instead.

Anyway, this is NGC4244, an edge on loose spiral galaxy in the constellation Canes Venatici. The processing was tougher than normal, it took several revisions of my image integration settings to get a good result against the noise. Generally I’m quite happy with the result.

Imaging Equipment

• Orion Optics 250mm F4.7 Reflector
• Baader MPCC
• Canon EOS 1000D (modified)
• Astronomic CLS CCD Filter

Guiding and Mount

• William Optics ZS66SD
• LX Modified Philips SPC900NC Webcam
• IR/UV Cut Filter
• Celestron CGEM Mount
• Guidemaster Autoguiding software

Exposures

• 27 x 5 minutes
• 15 x bias
• 15 x dark
• 15 x flat

The images were calibrated, aligned, stacked and processed in Pixinsight.

After spending Saturday night on NGC2403 we had another clear night on Sunday. I decided to make use of this weird good fortune and take another picture using the William Optics ZS66. Now I didn’t want to unrack the scopes, but I did want to guide so I made the decision to guide with the 250mm reflector, I figured I could use the focal reducer to give a wider field of view, but unfortunately could not get it to focus. No worries though, once pointed at the Pleiades guide stars were very easy to find.

It functioned pretty well as a guider, not that the field of view from the ZS66 makes it very critical. My major problem for the evening was dew, which the ZS66 dew shield could not hold off for long. Still it’s a reasonable result and I’m even happy with the processing. I also had a minor problem with the focuser on the ZS66 not coping well with the weight of the camera and sliding out. Locking it down caused the focus to shift. I ended up having to lock it down slightly in focus then slide it out carefully using my thumb as a lever to get focus. It’s a pain and I will probably take out an evening to dismantle the focuser assembly to see if it can be improved.

Imaging Equipment

• William Optics ZS66SD Refractor
• Wlliam Optics MkII Field Flattener
• Canon EOS 1000D (modified)
• Astronomic CLS CCD Filter

Guiding and Mount

• Custom Orion Optics 250mm Reflector
• LX Modified Philips SPC900NC Webcam
• IR/UV Cut Filter
• Celestron CGEM Mount

Exposures

• 31 x 6 minutes
• 21 x bias
• 21 x dark
• 21 x flat

The images were calibrated, aligned, stacked and processed in Pixinsight.

NGC-2403 is an intermediate spiral galaxy in the constellation Camelopardalis, and is apparently much tougher to image from my light polluted back garden than I expected it to be. It is however the first image taken with my newly rebuilt telecope, and that at least is a success. It was also a relief to reliable take 8 minute exposures and have the guiding mostly work as it should do. Even the meridian flip went smoothly.

The image itself however is not quite as good as I was hoping for, the out lying areas are lost, though you can just about see the spiral arms which is good. I’ll post more details on exposures numbers later on.