Alberto.I
Member
Hi,
I released a new image that I wanted to share here. It is my 2020 summer project and the target is the dust around the shark nebula. I knew that the shark nebula itself is quite dim, so I planned to go to a darker site than my second's residence backyard (supposed to be around SQM 21.5). I selected a hill at 30km from there, supposed to be SQM 21.87, at 1400m of altitude, and stayed for 2 nights. I gathered 8h of LRGB and then I returned to my backyard to continue imaging until a total of 24h of LRGB. To gather data was very easy now that I have my equipment properly "settled", but the processing was a totally different story. I needed 6 months until I became satisfied with the image.
This is the result:
Link to full resolution:
https://www.astrobin.com/43ckmd/0/
Technical Details:
Equipment:
Borg 101ED f/4 + QHYCCD 163M + HEQ 5 Rowan Belt Mod + AstroHub 2.0
Date:19/08/2020
Exposures:
Optolong B 36mm: 44x300" (gain: 174.00) -10C bin 1x1
Optolong G 36mm: 44x300" (gain: 174.00) -10C bin 1x1
Optolong L 36mm: 163x300" (gain: 174.00) -10C bin 1x1
Optolong R 36mm: 48x300" (gain: 174.00) -10C bin 1x1
Integration Time: 24.9 hours
Pixel Scale: 1,875 "/pixel
Adquisition Notes:
The adquisition was my first serious attempt imaging outdoors. I drove 30km from my second's residence (suppossed SQM21.5) until I found a place quite high (1400m) and supposed to be very dark (SQM21.87). The idea was to stay 2 nights gathering photons, but also testing my AstroHub 2.0 with a 100Ah LifePo4 battery and a 300W solar panel. Everything worked like a charm and now I know that I can rest for days or weeks with plenty of power. I gathered 8h of LRGB on this dark site, and the rest was taken from my backyard.
Processing Notes:
I needed 6 months to achieve the result I was looking for on this target. It has been a really challenging processing because the goal of the image wasn't the shark nebula itself but the surrounding dust. To pull out the dust was not a problem itsef, because the data was pretty clean. The real difficulty was the lack of contrast and to reduce the stars.
The processing of the LRGB non-linear stage has been quite simple using mainly curves and histogram transformations, no LHE, MMT, MLT or HDRMT enhacements at this stage.
The hard work was done previously, where I implemented new techniques that I think that worth to mention:
Multiscale Gradient Removal:
I had very bad gradients that didn't allowed me to push the data. I was struggling with this form months... and then I used this procedure, and fixed it in 1h. Simply perfect. In my opinion, it is the best gradient removal technique.
L to RGB starmatch:
L channel always has higher FWHM so stars grow when combining L with RGB. In this image was specially disturbing because the nebula, and the surrounding background is very faint so it needed and agressive star reduction. To try to avoid artifacts, I worked on the L channel before combining and I used a special contour mask produced by the difference between L stars and RGB ones, to apply erosion over the L stars. After this "L to RGB starmatch" there were 3 additional iterations of erosion, one of them specifically focused on small stars, another on medium stars, and a third overall to compensate a late black point adjustment.
Image Slicing:
Although no Starnet++ was applied to this image, it was used to create specific masks using the procedure that I published recently. For example to create an absolute starmask that I sliced in different star sizes to process them individually, or many other masks to isolate the DSOs involved in the image and be able to process them properly.
A discussion can be found here:
https://www.astrobin...inside-starnet/
...and a more in deep information of this procedure can be found in my website:
https://aiastro.word...de-starnet-wip/
-Refining procedure:
I used a procedure to refine the processing workflow using Process Containers. It allowed to change specific parameters of specific processes and see the influence in the final result in an easy and quick way. I'm just finishing a tutorial about it.
The following image shows some of the commented techniques. It is a comparison of the same processing only disabling the star erosion steps. It was a test made during the workflow refinement process.
You may notice that some of the stars (bigger ones) aren't reduced. This is because of the slicing of the starmask commented before. In first versions I tried to erode also this stars, but they ended by developing artifacts, so I isolated them and only worked on the medium/small stars.
Reducing the stars has been one of the most important aspect of this processing as it has been very agressive.
Hope you like it. Any comment/suggestion will be welcome!
Alberto.
I released a new image that I wanted to share here. It is my 2020 summer project and the target is the dust around the shark nebula. I knew that the shark nebula itself is quite dim, so I planned to go to a darker site than my second's residence backyard (supposed to be around SQM 21.5). I selected a hill at 30km from there, supposed to be SQM 21.87, at 1400m of altitude, and stayed for 2 nights. I gathered 8h of LRGB and then I returned to my backyard to continue imaging until a total of 24h of LRGB. To gather data was very easy now that I have my equipment properly "settled", but the processing was a totally different story. I needed 6 months until I became satisfied with the image.
This is the result:
Link to full resolution:
https://www.astrobin.com/43ckmd/0/
Technical Details:
Equipment:
Borg 101ED f/4 + QHYCCD 163M + HEQ 5 Rowan Belt Mod + AstroHub 2.0
Date:19/08/2020
Exposures:
Optolong B 36mm: 44x300" (gain: 174.00) -10C bin 1x1
Optolong G 36mm: 44x300" (gain: 174.00) -10C bin 1x1
Optolong L 36mm: 163x300" (gain: 174.00) -10C bin 1x1
Optolong R 36mm: 48x300" (gain: 174.00) -10C bin 1x1
Integration Time: 24.9 hours
Pixel Scale: 1,875 "/pixel
Adquisition Notes:
The adquisition was my first serious attempt imaging outdoors. I drove 30km from my second's residence (suppossed SQM21.5) until I found a place quite high (1400m) and supposed to be very dark (SQM21.87). The idea was to stay 2 nights gathering photons, but also testing my AstroHub 2.0 with a 100Ah LifePo4 battery and a 300W solar panel. Everything worked like a charm and now I know that I can rest for days or weeks with plenty of power. I gathered 8h of LRGB on this dark site, and the rest was taken from my backyard.
Processing Notes:
I needed 6 months to achieve the result I was looking for on this target. It has been a really challenging processing because the goal of the image wasn't the shark nebula itself but the surrounding dust. To pull out the dust was not a problem itsef, because the data was pretty clean. The real difficulty was the lack of contrast and to reduce the stars.
The processing of the LRGB non-linear stage has been quite simple using mainly curves and histogram transformations, no LHE, MMT, MLT or HDRMT enhacements at this stage.
The hard work was done previously, where I implemented new techniques that I think that worth to mention:
Multiscale Gradient Removal:
I had very bad gradients that didn't allowed me to push the data. I was struggling with this form months... and then I used this procedure, and fixed it in 1h. Simply perfect. In my opinion, it is the best gradient removal technique.
L to RGB starmatch:
L channel always has higher FWHM so stars grow when combining L with RGB. In this image was specially disturbing because the nebula, and the surrounding background is very faint so it needed and agressive star reduction. To try to avoid artifacts, I worked on the L channel before combining and I used a special contour mask produced by the difference between L stars and RGB ones, to apply erosion over the L stars. After this "L to RGB starmatch" there were 3 additional iterations of erosion, one of them specifically focused on small stars, another on medium stars, and a third overall to compensate a late black point adjustment.
Image Slicing:
Although no Starnet++ was applied to this image, it was used to create specific masks using the procedure that I published recently. For example to create an absolute starmask that I sliced in different star sizes to process them individually, or many other masks to isolate the DSOs involved in the image and be able to process them properly.
A discussion can be found here:
https://www.astrobin...inside-starnet/
...and a more in deep information of this procedure can be found in my website:
https://aiastro.word...de-starnet-wip/
-Refining procedure:
I used a procedure to refine the processing workflow using Process Containers. It allowed to change specific parameters of specific processes and see the influence in the final result in an easy and quick way. I'm just finishing a tutorial about it.
The following image shows some of the commented techniques. It is a comparison of the same processing only disabling the star erosion steps. It was a test made during the workflow refinement process.
You may notice that some of the stars (bigger ones) aren't reduced. This is because of the slicing of the starmask commented before. In first versions I tried to erode also this stars, but they ended by developing artifacts, so I isolated them and only worked on the medium/small stars.
Reducing the stars has been one of the most important aspect of this processing as it has been very agressive.
Hope you like it. Any comment/suggestion will be welcome!
Alberto.
