Pixinsight Photometric Colour Calibration fails - any ideas?

[Ouroboros]

Pixinsight photometric colour calibration fails on this image.  Anyone got any ideas why this should be so?  It also finds the wrong coordinates and focal length from the image.  However, it still fails when I correct these.  

If anyone would like a go at this I attach the xisf file below and the image details are as follows:

The image is the Pelican Nebula.  Focal length: 510mm  Pixel size: 3.76 microns

The correct coordinates of the centre of the image are: 

Center (RA, hms):20h 51m 44.079s

Center (Dec, dms):+44° 02' 43.645"

Date: 2021-09-04T21:20:13Z

The image has been calibrated with flats and darks and background normalised. 

Cheers.  

IC5070_dc_DBE_bn.xisf

[Budgie1]

I always use the Search Coordinates facility for the location and manually input the focal length and pixel size.

If it still fails at that then try changing the Database Server from the default of Strasbourg to one of the others, I normally use Victoria as a second option.

[Ouroboros]

Yes, thanks. I’ve tried manual input too. Oddly Pixinsight wrongly finds the focal length is 130mm from the image yet gets the pixel size right. Obviously I’ve corrected that and run it again. I have also tried the server at Cambridge UK and Victoria in Canada, but again no joy. Interestingly I ran a photometric CC on a calibrated image of the North America Nebula, right next door to the pelican, same camera and same telescope, and it got the PCC straight away, no problem. Mmmm? 

[teoria_del_big_bang]

Is this with OSC or mono & filters and if so what filters ?
Not that should stop it plate solving properly it should only affect the actual correction (I guess)

It fails for me when performing RANSAC is that the issue you had ?

Steve 

[Ouroboros]

19 minutes ago, teoria_del_big_bang said:

Is this with OSC or mono & filters and if so what filters ?
Not that should stop it plate solving properly it should only affect the actual correction (I guess)

It fails for me when performing RANSAC is that the issue you had ?

Steve 

Yes, it failed there. It’s an OSC, so I’m setting it as broadband. 

[teoria_del_big_bang]

Ah, just had a brainwave and remembered I had seen a similar issue on the Adam Block tutorials whereby the focal length and coordinates were wrong in the fits header so he changed them, just like you have done, in the parameters of PCC.
And, just like you it failed to platesolve.
He found the issue was that although you enter the data on the parameters the process still used the data from the fits header.
It only seems to use those parameters in the boxes if there is no values for them in the fits header, even if the focal length for instance is zero.
Which is not what I would expect I would expect the parameters to over-ride the header. 

So I first put the image on nova.astrometry.net to get correct coordinates.

Then used the fits header editor in PI to permanently change the coordinates and focal length.
NOTE: the process uses the values in degrees even though it displays them as H M S in the parameter boxes.

Then if I click acquire from image it gets the correct data from the image.

I am sure you are aware but to get good colour calibration when a very reddy image with lots of nebulosity use a preview of the background for background neutralisation, and also make the low limit much lower than 0.1 otherwise the sky colour comes out nowhere near the correct colour.

Then voila it platesolves okay.

Steve

Edited September 20, 2021 by teoria_del_big_bang

[Ouroboros]

@teoria_del_big_bang  Brilliant! Thanks. 😀 I’ll give it a go. First to find out how to access and edit the fits header. It’s an XISF file, but I guess that doesn’t matter. 

[teoria_del_big_bang]

2 minutes ago, Ouroboros said:

@teoria_del_big_bang  Brilliant! Thanks. 😀 I’ll give it a go. First to find out how to access and edit the fits header. It’s an XISF file, but I guess that doesn’t matter. 

No the same header, and more is in the xisf.

just make sure the correct image is selected then click on the part of the header to change and then change the text or value and click replace, do it for all things required and then I think you still need to press the blue square or drag the triangle over to the image for it to save them, it is a bit odd but you will work it out.

If you press acquire from image and the correct values appear then you know you have done it ok.

Steve

[Ouroboros]

1 hour ago, teoria_del_big_bang said:

No the same header, and more is in the xisf.

just make sure the correct image is selected then click on the part of the header to change and then change the text or value and click replace, do it for all things required and then I think you still need to press the blue square or drag the triangle over to the image for it to save them, it is a bit odd but you will work it out.

If you press acquire from image and the correct values appear then you know you have done it ok.

Steve

Bingo!  Yes, that worked fine.  I can carry on post processing now. Many thanks to you and your brainwave. 

Just for my and others’ future reference the FITSheader process is found in the Image menu. And for those with Warren A Keller’s Inside Pixinsight it is described from page 73 of his book. 

[teoria_del_big_bang]

42 minutes ago, Ouroboros said:

Bingo!  Yes, that worked fine.  I can carry on post processing now. Many thanks to you and your brainwave. 

Just for my and others’ future reference the FITSheader process is found in the Image menu. And for those with Warren A Keller’s Inside Pixinsight it is described from page 73 of his book. 

Well some things stick in there but most bits of learning seem to go right through these days 🤣