StasV
Well-known member
Hi All!
For stars accurate processing it is necessary to separate the star mask into the several masks in accordance with the size of the stars (see Fig. 1). From now on, I will call this as “multi-caliber star masks”.
This procedure also be useful when you process stars and nebula separately using StarXTerminator / Starnet++. Make a multi-caliber star masks are a difficult painstaking work. It is described in these videos by Adam Block
video1
video2
The method is based on iterative application of minimum / maximum filters. This method has the following disadvantages:
1. Difficult and long
2. The shape of the masks is degrading
3. The Star Diffraction spikes go into the mask of small stars, and the star’s centers into the mask of large stars (Fig. 2).
I have a suggestion, to create a module that will automatically separate the initial mask into multi-caliber star masks. The idea is not to work with the entire image using graphical methods, but to find all the stars on the initial mask. Isolate them. And work with each star as with an object. Analyzing size and shape. This approach opens up many possibilities:
1. Precise separate of masks by size into any number of masks.
2. Separation out of non-stellar objects. For example, small elongated galaxies using the shape criterion.
3. Apply different smoothing amount and expansion for different star calibers.
You can go even further. And try to develop an algorithm for separating not masks, but the stars themselves. The idea is roughly shown in Fig. 3. Perhaps Russell Croman ( StarXTerminator author) will be interested.
For stars accurate processing it is necessary to separate the star mask into the several masks in accordance with the size of the stars (see Fig. 1). From now on, I will call this as “multi-caliber star masks”.
This procedure also be useful when you process stars and nebula separately using StarXTerminator / Starnet++. Make a multi-caliber star masks are a difficult painstaking work. It is described in these videos by Adam Block
video1
video2
The method is based on iterative application of minimum / maximum filters. This method has the following disadvantages:
1. Difficult and long
2. The shape of the masks is degrading
3. The Star Diffraction spikes go into the mask of small stars, and the star’s centers into the mask of large stars (Fig. 2).
I have a suggestion, to create a module that will automatically separate the initial mask into multi-caliber star masks. The idea is not to work with the entire image using graphical methods, but to find all the stars on the initial mask. Isolate them. And work with each star as with an object. Analyzing size and shape. This approach opens up many possibilities:
1. Precise separate of masks by size into any number of masks.
2. Separation out of non-stellar objects. For example, small elongated galaxies using the shape criterion.
3. Apply different smoothing amount and expansion for different star calibers.
You can go even further. And try to develop an algorithm for separating not masks, but the stars themselves. The idea is roughly shown in Fig. 3. Perhaps Russell Croman ( StarXTerminator author) will be interested.