Author Topic: A Batchpreprocessing issue...  (Read 2438 times)

Offline Rmoore7167

  • Newcomer
  • Posts: 7
A Batchpreprocessing issue...
« on: 2018 September 10 16:15:51 »
Hi
Im really new to this and have been Harrys flow for batch processing and have been trying to do some recalibration but keep getting a failure notice that states "ImageCalibration" cannot execute instance in the global context. Reason: No master calibration frames or overscan regions have been specified"

I don't have any masters and made sure the the Use mater bias, dark and flats are unchecked..But when i get failure notice I see that bias and darks are now checked.  Under global options I have selected ; Optimize Dark Frames, Generated Rejection maps, save process log, up-bottom FITS.

any help would be appreciated....probably something simple.  I have

Bias- 18 files
Darks- 18 files (1x1)
Lum- 8 files (1x1)
red 8 files (2x2)
Blue 7 files (2x2)- won't work as need one more
green 8 files (2x2)
HA 8 files (2x2)

Not sure if thats some of the issues or not so thought Id post.

Thanks...let me know if I need to communicate more

Offline ngc1535

  • PixInsight Old Hand
  • ****
  • Posts: 326
Re: A Batchpreprocessing issue...
« Reply #1 on: 2018 September 10 16:31:23 »
Hi,

Just looking at your list...
I see you have data binned 2x2 but calibration files at 1x1. You need to match binning with regards to darks and biases.
Additionally 7 frames is of course just fine to combine... more frames will give you better statistics of course.

Please let me know if I am on the right track.
Thanks,
Adam

Offline Rmoore7167

  • Newcomer
  • Posts: 7
Re: A Batchpreprocessing issue...
« Reply #2 on: 2018 September 10 16:39:23 »
Thanks....just to be clear. My bias, darks and Lums are all 1x1.  BUT....I have my RGBHA all at 2x2.  I was told that to take my lights as 1x1 everything else at 2x2....Is this incorrect?

Offline ngc1535

  • PixInsight Old Hand
  • ****
  • Posts: 326
Re: A Batchpreprocessing issue...
« Reply #3 on: 2018 September 10 22:23:25 »
In order to calibrate data you take at 1x1 you have to have darks and biases that are also 1x1.
If your data is 2x2 you need to take biases and darks at 2x2 as well.

Flats are a little different. Technically you can resample flats (resize them) and use them if done properly. However, to keep the theme going, you can be safe and match the binning for the flats as well- then you will not have problems when using BPP.

So the statement "Take my lights at 1x1 and everything else at 2x2" is incorrect if we are talking about the calibration data (biases and darks). These you match.

I suspect what you are thinking of the creation of LRGB (or LRGBHa) type images for which, yes, one method is to bin the color information in order to minimize read noise.
However, this doesn't deal with your initial problem/error for BPP which is concerned with matching biases, darks, and flats to your images. There needs to be a 1 to 1 correspondence in binning. Since you do not have any binned 2x2 darks or biases... BPP says it can't calibrate your data.

-Adam

Offline stevek

  • PixInsight Addict
  • ***
  • Posts: 148
Re: A Batchpreprocessing issue...
« Reply #4 on: 2018 September 11 08:35:14 »
It's painful to have taken many quality lights at 1x1 then done something stupid like taken the flats at 2x2...... :(  I have some data sets like that.  Just can't bring myself to delete them and have hung on to them in the hope that when my skills have improved I might be able to do something with them....

Offline ngc1535

  • PixInsight Old Hand
  • ****
  • Posts: 326
Re: A Batchpreprocessing issue...
« Reply #5 on: 2018 September 11 09:12:48 »
Downsampling flats will work just fine with IntegerResample. Upsampling obviously will not recover pixel to pixel variation that may be recorded in your 1x1 data. However, an upsampled flat will certainly work to take care of vignetting and dust motes- the large scale stuff that concerns astrophotographers. So if you have 2x2 flats you can certainly use them. There is an issue of maintaining flux when you upsample- but I suspect it will all work out. You would upsample with IntegerResample and if you are using BPP you will need to add the flats in "custom" in order to override the FITs header binning value of "2" and force it to be "1".  Then you are off to the races.
-adam

Offline stevek

  • PixInsight Addict
  • ***
  • Posts: 148
Re: A Batchpreprocessing issue...
« Reply #6 on: 2018 September 11 10:27:51 »
Wow, so all I need to do is custom in the BPP script Adam?  Or do I need to use the IntegerResample and a prior process?
I can recover about 80 hours of M81/M82 wide-field data if I can do this!!!!! :)
-thanks, Steve

Offline ngc1535

  • PixInsight Old Hand
  • ****
  • Posts: 326
Re: A Batchpreprocessing issue...
« Reply #7 on: 2018 September 11 12:16:13 »
Resize the flats using IntegerResample...  (probably with an image container) then add them using Custom in BPP.  BPP will not resample the flats for you.
Please report back and let us know if it works for you.
-adam

Offline stevek

  • PixInsight Addict
  • ***
  • Posts: 148
Re: A Batchpreprocessing issue...
« Reply #8 on: 2018 September 12 00:46:30 »
I Will do Adam, working overseas currently so will check it out in a few days when I'm home.
- Steve

Offline stevek

  • PixInsight Addict
  • ***
  • Posts: 148
Re: A Batchpreprocessing issue...
« Reply #9 on: 2018 September 14 00:23:21 »
Worked perfectly!
 :) :) :) :) :)