Jump to content


Photo

Image Processing Routine

image processing routine

  • Please log in to reply
4 replies to this topic

#1 Neil

Neil

    Administrator

  • Administrators
  • 1,493 posts
  • LocationThe Netherlands

Posted 09 December 2012 - 02:27 PM

OK, you found your target and managed to gather some data, now how should you proceed to process your image? Well firstly you need to establish a routine to get your newly captured data in the best possible condition for processing into that beautiful image that you have seen in the lastest astronomy magazine.

 

Here is a typical step by step imaging route, post image capture.

 

1. Image calibration, part of the image that you captured is unwanted noise.  The unwanted noise has many sources that will be discussed as an advanced topic later. At a bear minimum you need to be shooting  BIAS frames and DARK frames. These are then used to cancel out the majority of the camera's noise. To improve image quality further, you should also shoot FLAT frames to record optical defects such as dust motes and vignetting.

 

Calibration frames

  • BIAS
  • DARKS
  • FLATS

 

2. Once your LIGHTS have been calibrated they should look a lot cleaner, most of the noise will have been removed along with the HOT & COLD pixels. Now it's time to stack the frames to produce a MASTER frame. Depending upon how many frames that you managed to gather you have several choices to combine them statistically.

  • Sum
  • Mean
  • Median
  • Some form or other of Sigma clipping

With a limited number of frames, lets say less than 4, Sum is your best option, up to 6 frames MEAN and MEDIAN come in to play and with 10 or more frames SIGMA CLIP can work its magic. Sigma clipping is a way to further eliminate noise or cosmic ray hit or even satellite tracks. Sigma Clipping only accepts data that is present in all of the frames, everything else is removed.

 

3. Now that you have your stacked reference image it's time to see just how good it is? Using your favorite image processing tool, stretch the image to reveal all of the data that you have captured. Be very careful here!!! Do not clip the black point as this will kill your background making your stars look over processed and saturated.

 

Black Point.png

 

Here you can see that there is still some space on the far left of the histogram.

 

4. Now that your masters are stretched it's time to check for the dreaded gradient! Gradients appear for all sorts of reasons, mostly light pollution is the main culprit. Some software tools are better than others at removing them. However you plan to tackle them, you need to do it at this stage.

 

5. Combine your nice MASTER images into a colour image using the software tool of your choice.

 

6. Colour balance the image to ensure a natural look for all of the components: Stars, Nebula and background etc...

 

7. Perform as many tweaks or enhancements as you like.

 

I will create an advance routine shortly with more images depicting the various steps in the process.

 

However, I have just learnt that Sir Patrick Moore has past away peacefully at his home in Selsey, West Sussex this afternoon.

 

RIP Sir Patrick Moore, an inspiration to many young and now older astronomers alike. 

 

Clear but very sad skies

 

 

Neil.



#2 jlake

jlake

    Advanced Member

  • Moderators
  • 106 posts
  • LocationPomfret, CT

Posted 09 December 2012 - 07:54 PM

Using PixInsight, I find it very useful to just screen stretch the image (with Screen Transfer Function) and do several steps while the data is still linear. Deconvolution, noise reduction, and dynamic background extraction are all better when done on linear 'deep' data, yes? And for color data, background neutralization and color calibration make much more sense in linear, non-stretched space.



#3 Neil

Neil

    Administrator

  • Administrators
  • 1,493 posts
  • LocationThe Netherlands

Posted 10 December 2012 - 11:57 AM

Using PixInsight, I find it very useful to just screen stretch the image (with Screen Transfer Function) and do several steps while the data is still linear. Deconvolution, noise reduction, and dynamic background extraction are all better when done on linear 'deep' data, yes? And for color data, background neutralization and color calibration make much more sense in linear, non-stretched space.


Hi Josh,

I fully agree with you. I have just made a PI video that I'm trying to voicing over at the moment for submission here, however, if you want to pick this subject up I would really appreciate it. Either a video or a post with screen shots should some of the PI processes etc...

Regards


Neil.

#4 Donna111

Donna111

    Newbie

  • Members
  • Pip
  • 4 posts

Posted 23 December 2013 - 03:05 AM

HI there

I am a beginner of image processing.And i want to get a powerful image processing program which supports to edit image directly offline.I know there are many  third party program for image processing.But i want to

get a free trial package before i decided to purchase it.Thanks for any suggestions.



#5 Tinman

Tinman

    Member

  • Members
  • PipPip
  • 95 posts
  • LocationCheyenne, WY. USA

Posted 28 September 2015 - 03:01 PM

I use Maxim DL, CCD Stack 2, and Photoshop CS6 for  processing.  CCDOps for bloom removal.

I am trying to "get a handle " on narrow band processing, and keep having problems with getting it to come out right.

Any ideas?

 

Rob C.


Rob C.

Cheyenne, WY. USA

 

Pier mounted AT6RC with SBIG ST10-XME/CFW8a/ AO-7, WarpCorp driven

Pier mounted LX200 10"  with SBIG ST7-E/CFW8/ AO-7, with Optec TCF-S focuser.





Thank you for your interest in our Astronomy Forum...
It looks like you haven't joined us yet?
Why don't you register to get all the features of being a member! Its free!