PIXINSIGHT TUTORIAL: Elimination of Bright Star Halos (English)

Author:  Rafael Rodríguez Morales              Level: Advanced

Hello again, this time and within the tutorials on processing astronomical photographs with PixInSight I present this tutorial that explains the processing to eliminate halos in very bright stars. This effect or optical aberration is normally caused by the optical train (filters, reducers,), leaving the only option to eliminate or reduce it in the processing phase of the final photograph. The method that I present here I have developed and complemented based on ideas , being the original method, being quite practical, simple and effective.

The PixelMath tool will be used  as the main tool, which allows to materialize mathematical calculations with each pixel of a photograph 

The object is to create a mask that perfectly covers the circle of the halo, for this we will use the mathematical formula of the circle:

iif (sqrt ((x-CX) ^ 2 + (y-CY) ^ 2) <R, 1,0

Conditional formula ( iff ) that has 3 constants to be defined manually and that defines the size and position of the circle within the frame of the photograph .:

 R = Radius of the circle

CX = Center of the circle, X coordinate

CY = Center of circle, Y coordinate

We will use the attached image, in which a large halo is observed around the star Antares.

We calculate the position of the center of Antares, for this we use the Dynamiccrop tool  . The coordinates start in the upper left corner, we can know that the center of Antares is at the coordinate X = 1840, Y = 2336 approximately.

In the same way we can know the approximate radius of the circle, R = 863.

Defining the parameters and formula in PixelMath (in this case we define R = 1000 to see it better).

Applying PixelMath we create the mask that we apply to the photograph.

Near vision and the position of the mask in relation to the halo. You will need fine adjustments to the position and radius of the circle. You can try on and create the necessary masks until you see that it fits the halo.

Once set it is convenient feather the circle  ut ilizando  MultiscaleLinearTransform.

Applying the mask and blurring the edge as precisely as possible, we use  CurveTransformation to reduce the transition of the edge of the halo. We cancel layers 1 to 6.

.

Final version.

And this is it ... Thank you

email:  rrodrigum@gmail.com