ImageIntegration parameters

[kwiechen]

Hello,

I have used PI for several months to process images (and now I feel that I have wasted time and money with PS). Stacking was done with MaximDL with varying results. Now I have tried PI's 'stacking' modules (ImageCalibration, ImageIntegration and StarAlignment) for the first time with much better result using the options described in http://pixinsight.com/tutorials/master-frames/en.html and http://www.harrysastroshed.com/pixuser/intergration.html ( Thank you for these excellent docs!).

Is it useful or necessary to use different ImageIntegration parameters for different filters/binning to get the best results?

Best,

Kai



 

[sreilly]

As a rule of thumb I use the following for deciding which rejection algorithm to use:

3-5 Images      Percentile Clipping or
                      Average Sigma Clipping

5-10 Images    Sigma or
                      Average Sigma

10-20 Images  Winsorized Sigma Clipping or
                      Linear Fit Clipping

20+ Images    Linear Fit Clipping

As for Combination method, depends on the data sets. If I have say a satellite trail in one or several images but the rest of the data is really good, I'll use Median. Saying that, maybe if I fooled around with the Pi\xel Rejection (2) settings I might get these trails removed but I haven't seen that. Otherwise I'd use the Average method. I tend to use large sets of data and other than visually I can't say I've seen any difference in the two methods other thaen when I do have something like satellite trails to remove. But also keep in mind that many of my images will have 16 or more hours of total data being use usually under 2.5 arc seconds or better.

[kwiechen]

Thank you for the explanation! The background is that I am playing a bit with a simple batch script to speed up this process. If I correctly understood, it is OK to use parameter sets for generation of master bias, dark and flat and it is necessary to be more flexible when combining the light frames. If possible I try to use > 20 * L with 1 * binning and RGB  2 * binned with > 10 frames per channel. So I can try to end the script with the light frames calibrated and aligned sorted by filter.

Best,

Kai

[mmirot]

Somewhere in the past Juan or Vicent had posted the best setting for light frames , flats , darks and bias.

If you come up with a good script don't forget to share it with the group !


Easy use in calibration is one area that PI needs some help

Max

[kwiechen]

I have tried to write a simple prototype script to speed up the process of calibrating and aligning light frames. This is my first attempt with javascript; so there is a lot of work to do.

The script is written to support my workflow: FITS data from a monochrome CCD, binning 1 or 2. The script puts the light frames calibrated and aligned to folders sorted by filter information in the FITS header. So I can easily pick up the calibrated light frames to optimize the rejection parameters for ImageIntegration. Some FITS data is written back to the master calibration frames. So it is possible to feed the script with masters previously created.

Best

Kai

 

[mmirot]

I ran you script stops at line 515 NumericControl is not defined.
Looks interesting. Please post any changes


Max

[kwiechen]

Unfortunately the include-Statement for NumericControl was lost - can you please try this version?

Best,

Kai

[mmirot]

Sure I as soon as I get home.
Max

[mmirot]

By the way a calibration, alignment and integration script was started last year called the preprocess pipeline.

It has some good features but it took a long time to setup.
Also the script setup was not the easy to understand.

If your interested no one has done anything on it in about six months or more.

http://pixinsight.com/forum/index.php?topic=2861.0

Max

[kwiechen]

I have just started to use PI to stack my images. My very simple batch script was written to speed up this and to better understand the processes. The script uses only parameters I am using for now and so it it not very flexible. But if you have frames from a monochrome camera with XBINNING, FILTER and EXPTIME keywords set it may work for you.

Best

Kai

[mmirot]

http://pixinsight.com/forum/index.php?topic=3651.0
Now we have official documentation on the subject !

[mmirot]

Kai,

 There is no paths to select for prior created masters or is this built in somehow?
 It would be nice have version that lets you use previously created masters.

Btw I have a monochrome camera and the Fits key words work so far.

Max

[kwiechen]

Hi Max,

if the masters are created with the script just put them via add files into the correct tabs and check the appropriate boxes. If you have masters previously created with another program be sure to have the FILTER and EXPTIME keywords set. I have added an update of the script that handles dark frames with EXPTIME differerent from the light frames.

Best,

Kai

 

[mmirot]

Sounds good. I will playing a bit more with it in next few days.
Keep making improvements.  It would be great if you add a finial integration step.

Max

[kwiechen]

The code for the final integration step is commented out for now because I have found different integration parameters useful for combination of luminance and RGB frames - but I think about this. You can try to reactivate this - it will generate stacked L, R, G, B images in the master frame directory.

Best,

Kai