Jump to content

Banner.jpg.b89429c566825f6ab32bcafbada449c9.jpg

ASTAP v0.9.9 with new internal alignment routine for FITS file stacking


han59

Recommended Posts

11 hours ago, han59 said:

At what resolution are you imaging? Modern digital cameras have a large sensor of 12 or 20 megapixels which could result in oversampling and long processing time.  Binning the raw in the beginning of the stacking process could improve the processing speed and the signal to noise ratio.

I'm imaging at full res, unfortunately the t5i doesn't have the option to reduce raw resolution, only jpeg. I'll try again astap again tonight and let you know how it goes. 

Thanks a lot. 

Link to comment
Share on other sites

  • Replies 34
  • Created
  • Last Reply

Looking to the 5Ti specification is indicates RAW: Approx. 17.90 Megapixels (5,184 x 3,456)  fixed size. This is no problem, but it will be a little slow. So first try only a few files.

I just released ASTAP version 0.9.82 In this version I added a tool to split the RAW (FITS) in four seperate files of red, green, green and blue. Some call this "Extract Bayer Plane". You get 4 files (R,G,G,B) of half resolution.  Nise should then reduce. Maybe the stacking speed has increased but you also have four files rather then  one file. It is a new path but not fully tested.. T omorrow I will test the option for stacking an old set.

The orginal idea was to add this tool voor small-band filter imaging. For example in h-alpha you use only the red channel reducing the total noise.

Best regards, Han

 

 

Link to comment
Share on other sites

Current the best approach is to use the converted files in the original RAW resolution 5184 x 3456 (grey image)  The program will apply the darks and flat & bias correctly in the same 5184 x 3456 dimensions, convert them to color using so called "Bilinear interpolation" as used in DSLR camera's maintaining the 5184 x 3456 dimensions and finally stack the coloured images. So it is handeling 3x 5184 x 3456 pixel arrays.

My idea of separating the R, G, B, pixels into 5184/2 x 3456/2 images will only work if your applying red-darks, green1-darks , green2-darks, blue-darks made in the same way. That can be done but requires four manual stacking process steps and a final step to add the four channels together.

That is not efficient.  Thinking about it, I assume now that this all can be realized internally and more efficient in ASTAP.  Rather the applying the time consuming "Bilinear interpolation" algorithm, in the end I can just simply add the red, green, blue pixels together and create an half sized image. That will reduce the noise since it will not use neighbouring pixels in the calculation and increase the stacking speed.  If the image is near or over-sampled (stars 4x4 pixel large or more) this looks the way to go.

I will try to implement it as a new option. At the moment I'm imaging with a mono camera and have only one DSLR image test set. If you could provide some test material that would be nice.

Han

Typical Bayer CFA:

bayer_cfa.png.8b5d16ebbbe983707fee606b06d6dfdd.png

 

Link to comment
Share on other sites

Hi Hi Han,

I did two separate stacks on Pixinsight and ASTAP to compare results (for fun). 
As I previously mentioned, I tried imaging in mixed BINS (HA bin x1, OIII bin x2).

ASTAP, initially did not want to stack, as I placed x2 and x1 bin BIAS and Flat files for stacking. so I stacked master Flats Bin x1 separately from x2.

The final RGB result was quite funny, it was something like .GIF, flickering, and 3D image at the same time.

image.thumb.png.c5086c55abbae1d088ec2f36f725ec16.png

Print screen is not able to reflect what was going on.

However, 

Once I took separate HA and OIII stacks and combined them I got a result below :)
1638052092_VeilNebula-ASTAP3.thumb.jpg.f947dfe96cd2bf718e0f3be1c946ca05.jpg
 

In comparison, I also stacked and combined the same images in PI and the result was.....

914092221_VeilNebula.thumb.jpg.16a13ef29fc981f1c92c416521817749.jpg

I felt that noise control was much easier on ASTAP files and the final image is a bit better :)

ASTAP left temp .ASTAP_Solution files and had to delete them manually.

L_VeilN_OIII_2018-07-11_23-16-19_Bin2x2_250s__-20C_ 90xFD 100xF 19xD 24xB_stacked_ 32x300R 0x0G 24x250B 0x0RGB 0x0L _stacked.txt

Link to comment
Share on other sites

That is an impressive result. The images look very sharp and not over-sampled. Still I will try to implement the "bayer_binning" (I have not better  name) to see what it does to the noise or for case the images are over-sampled.

Link to comment
Share on other sites

Yep mate, I feel happy! :) 
It was my first proper try with my ASI after the arrival of that bloody adaptor for my coma corrector ;)

I am still quite a newbie in this field, - but the result impressed me much! :) 
It is also my first time I am actually waiting for the Autumn and for the Longer Nights :) 

Link to comment
Share on other sites

The option I proposed above for DSLR cameras/OSC images is in DSS and PI called "Super-pixel".  It is now implemented in ASTAP v0.9.83. The stacking speed is 3 times faster! The final image is half size of the original. I have no idea what is does with the SNR value but assume it could be beneficial for nebula having over-sampled images.  Couldn't find any performance report for this method.

If your using a H-alpha filter in front of an OSC camera, by using the super-pixel de-mosaic method, you could just remove the green and blue signal to create a noise free red image and then convert it to mono. This makes processing simpler. There is no interpolation between the R,G,G and B pixels, so any noise from G an B pixels can be separated from the red.

Han

super_pixel.thumb.png.c03c94dc19b048d5c15b46c7d393f755.png

 

Link to comment
Share on other sites

The de-mosaic routine in ASTAP version 0.9.83 has a bug resulting in a run-time error. Please download version 0.9.84.

This version has also a new one star alignment routine intended for moving objects like a comet. Below as an example. A stack  of 19 images of 50 seconds centered on the  comet  Komeet 21P / Giacobini - Zinner.

Han

Update:

  Version 0.9.84 has bug in the sigma-clip routine causing a run time error. This has been fixed in 0.9.85

 

 

61515586_21P_Giacobini-Zinner_50s_20180717_224143_0x0R0x0G0x0B0x0RGB19x50L_stacked.jpg.493e6a91e85bec428ac1329c51f7a6b2.jpg

 

 

 

Link to comment
Share on other sites

  • 4 weeks later...

The development continuous. The latest ASTAP version 0.9.92 has a blink routine (for FITS files) to detect minor planets or nova or just make a video of a comet moving:

Youtube demonstration:

https://youtu.be/7cheLI-3ggQ

At the moment the blink sequence can't be exported , but you could capture it by a screen capture program.

So ASTAP has become more a software suite. The main functionalities are:

  • - FITS viewer and manipulator including annotation of deep sky objects and stars.
  • - Stacking program, blinker
  • - Plate solver with two types of command line interfaces, remote accessible.

Feedback is welcome

Han

 

The new blink tab:

astap_blink_routine.thumb.png.7bc48683cabb7ee23126db53df1ed688.png

Link to comment
Share on other sites

Archived

This topic is now archived and is closed to further replies.

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...

Important Information

We have placed cookies on your device to help make this website better. You can adjust your cookie settings, otherwise we'll assume you're okay to continue. By using this site, you agree to our Terms of Use.