The Northern Milky Way with Cygnus, Lyra, Aquilla and Delphinus.

In late August 2019, I spent a few days at an ‘Astronomy Week’ on the Dorset coast and, miraculously, had three clear nights.  The first two of these were spent showing the group Jupiter and Saturn, low in the south, and some of the other interesting objects in the heavens which, happily, had little light pollution to contend with.  On the third night, I demonstrated some aspects of astroimaging but taking less time imaging individual targets than I would normally do.  Thus the resulting images were initially not too impressive and required quite a bit of processing in Adobe Photoshop to produce an acceptable image.  One could also use the free program GIMP and there are a couple of articles in the digest about its use.  This article demonstrates the rather complex processing that I employed to produce an image of the northern Milky Way – high overhead in August.

The first image below is the final result − but with 1/5th the overall dimensions.  The second shows the ‘Summer Triangle.  A gap has been left in the line from Altair up to Vega to show ‘the coathanger’ – a nice asterism.  The third highlights the constellations Cygnus, Lyra, Aquilla and Delphinus.  Though not visible at the scale of these images, it was pleasing to see that the two stars making up the double star Albireo in Gygnus were individually captured and that the pair of the ‘Double Double’ star in Lyra were also resolved.

The individual frames were taken using the Samyang 12mm lens mounted on my Sony A5000 camera.  This gives a 35mm full frame equivalent focal length of 18mm. The digest includes articles about both the camera and lens along with one about star trackers including the StarSync tracker that was used to mount the camera.   A total of 25, 30 second, exposures were taken with an aperture of f/3.5 and an ISO of 200.  To be honest, this total exposure is not really enough to bring out the Milky Way well.  The image below shows the result of ‘stretching’ an individual, 30 second, frame – which does show the Milky Way but is, of course, very noisy.  [The article in the digest ‘What ISO to use for Astrophotography’ suggests that if the camera is ‘ISO invariant’, as is the Sony A5000, then using a lower ISO may actually be better than using ISO 800 or 1600.]

The 25 frames were stacked in Sequator (see article in digest) to give an effective exposure of 12.5 minutes.  The band of the Milky Way is not visible in the stacked image without stretching.

The only way I found to successfully process the resultant image was to first separate it into two images, one of the stars alone and one of the band of the Milky Way having no stars. These were then processed individually before combining them into the final image seen above.

Splitting the stacked image

The stacked image was loaded into Adobe Photoshop and the ‘Noise and Scratches’ filter applied with a radius of 15 pixels.  The filter ‘thinks’ the stars are dust and removes them leaving just the band of the (barely visible) Milky Way.  This was saved as ‘Milky Way band’.

This image of the Milky Way band was selected and copied (Ctrl A, Ctrl C).  The original stacked image was loaded for a second time and the Milky Way band image pasted over it (Ctrl V) giving two layers.  The ‘Difference’ blending mode was chosen and the two layers flattened to give an image of the stars alone.  (((Stars+Milky Way band) – Milky Way band) = Stars).

Processing the Milky Way image

The image brightness was stretched using the ‘Levels’ command with the middle slider moved to the left to the value of 1.2.  This brightens the whole image but the fainter parts more so.  A second application of the Levels command was then used to bring up the ‘Black Point’ of the image by moving the left slider up to the start of the histogram.  The procedure was repeated until the band of the Milky Way showed up well. 

To increase the local contrast in the image, the ‘Unsharp Mask’ filter was applied with a radius (its maximum) of 250 pixels and amount of 30%. 

As there was not a sufficiently long total exposure to give a ‘smooth’ result, the ‘Gaussian Blur’ filter was applied with a radius of 15 pixels.  This resulted in the ‘Milky Way band’ image which was to become the basis of the final image.

Processing the Stars image

There is a fundamental problem with producing an image of a star field that approximates that seen by the eye in that many of the brighter stars will all appear at a similar brightness when the image is stretched to bring out the fainter stars and so the constellation patterns will not be apparent. The method described below attempts to give an image that more closely matches that seen by the eye. 

The stars image is saved as two separate images, Stars-A and Stars-B to give two copies.   The Stars-A image was stretched using a few applications of the Levels command with the central slider moved to the left from 1 to 1.2.  Each application increases the brightness of the fainter stars giving, after 4 applications of the filter, the result below which was saved as Stars-A-Stretched. 

The Gaussian Blur filter was applied to the Stars-B image with a radius of 5 pixels.  This tends to remove the fainter stars and reduce the brighter ones in brightness (but which are now somewhat larger in diameter).  Their brightness is brought back using the ‘Brightness/Contrast’ tool.  Several applications of this procedure will produce an image of just the brighter stars whose size has been increased.  As in star atlases where brighter stars are shown as bigger disks, this helps to make them stand out producing the ‘Bright Stars image’.  To enhance the star’s colour, their saturation can be increased using the ‘Hue/ Saturation’ tool giving the ‘Bright Stars enhanced’ image.

Combining the three images into the final image.

The ‘Stars-A-Stretched’ is loaded back into Photoshop, selected (Ctrl, A) and copied (Ctrl, C).  The ‘Milky Way band’ image is loaded and the ‘Stars-A-Stretched’ image pasted on top (Ctrl V) to give two layers which are flattened using the ‘Screen’ blending mode.  This adds the stars into the image. 

Finally, the ‘Bright Stars enhanced’ image is loaded, copied and pasted over the just flattened image, again flattening the two using the ‘Screen’ blending mode.  This has then combined all three images to give the final result as shown at the top of the article.

Considering the limited exposure time, the final result was quite acceptable and the Samyang Lens, stopped down from f/2 to f/3.5 acquitted itself well as did the tracking mount so, together, providing pin-point stellar images.