Collecting  Photons


Star Wars III

Fighting the bright stars in the Pleiades

 

Two parts mosaic of the M45 cluster. ASA 10" astrograph, Modified Canon 350D, total integration 40x8' + 17x30" for each panel in three nights. Calibration in Iris, processing in PixInsight and Photoshop CS. 

My intent during the shooting and processing of this image was not to produce a very deep rendition, but rather to enhance the subtle interplay of colours of the nebula and surrounding stars. One of the early problems encountered during the acquisition of the "deep" data set (40x8') was that the bright stars of the cluster overpowered the surrounding area, badly saturating, being way too obtrusive and loosing any trace of colour. To overcome this problem I acquired a second data set formed by a sequence (17 subframes) shot at a much shorter exposure: only 30 s. The simple procedure presented here shows how is possible to use this "light" data set to seamlessly correct the brightest areas of the "deep" data set.

Back Home


1) The two data sets

After calibration, registration, correction of background, and rough colour balance, the two data sets received a preliminary stretch by means of the hyperbolic arcosine function implemented in Iris. To enhance the fainter nebula, the "deep" data set received a more pronounced stretch (asinh 0.003 12) than the "light" set (asinh 0.001 6).  


2) Preparation of the star mask

Now, the task is to blend the two data set without causing any artefact at the edge of the fused areas. The key for a correct blending is a mask that allows only the bright stars providing an intermediated area properly feathered for the fusion. The mask is prepared in two steps: first, from a luminance layer (an actual luminance shot though a clear filter, or an artificial luminance synthesised from the RGB channels) a ne image is prepared by drastically regulating the levels in order to maintain only the upper 20% of the histogram. Little nebula is present here and only the brightest stars. Care has to be exercised to determine the proper threshold point. Then this image is low pass filtered through a Gaussian blur operator (3.6 pixels in this case). Levels are further adjusted to attain the right transparency.

 

 


3) Final blending

The blending is realised in Photoshop CS by creating two layers. The background is formed by the "deep" image, while the upper layer if formed by the "light" image. The mask is applied to the upper layer: the white areas of the mask make the upper layer visible, while it is invisible in correspondence of the black area. At this stage the final adjustment of the mask can be applied, having an immediate feedback on the end result.

The light layer is applied with a transparency of 80% to further help the smooth transition.


 

4) End result

Rolling over the mouse with show alternatively the image before and after the process. Not only the brightest stars are reduced in size, but the colours info, that are present at the unsaturated edges of the star image, are brought closer to the star center. This fact will help further processing directed toward increasing the star chroma.

 

Roll over with the mouse to see a blinking comparison of the images before and after processing.

 

Send anything that comes to mind to Gimmi Ratto gimmi@in.cnr.it

Copyright © 2009 by Gimmi Ratto. (January 8, 2009)