Having extreme difficulty getting colour balance with one image

[robertm]

Hi All,

I'm usually fairly self sufficient when it comes to PI having followed the excellent tutorials on the web but this image is getting the better of me.  I took this at Kelling Heath recently, the image is 4hrs of exposure of NGC1848 at f/2 taken with an SXVF-M25C one shot colour.  Another image I took during the same session of NGC1396 worked out very well and I'm following the same process but after a week of trying it's really giving me a headache !  Colour balance seems all wrong and star colour is absent.

My process flow has been.

1) Calibrate/Stack in MaximDL then over to pixinsight
2) ABE.
3) Background Neutralization
4) Colour Calibration

The last two steps are where it seems to go wrong.  There is so much nebulosity that almost no sky limited areas exist across the whole image.  I've tried a number of areas but to no avail and the image still looks pink with no star colour.  I'm sure there's something fundamental I'm doing wrong but for the life of me I can't see it.  Could it be the ABE that's not right ?

This my effort on flickr:



This is a link to the full res image: http://farm9.staticflickr.com/8321/8016542318_8888c842cc_o.jpg

Any pointers or help would be greatly appreciated.

Unfortunately even in bin 4x4 8 bit mode the file is still too large to be uploaded.

Many thanks in advance.
Robert

[Geoff]

I had a similar problem with the eta Carinae nebula. I found that lining up the peaks of the R, G and B in histogram stretch worked pretty well.See this thread
http://pixinsight.com/forum/index.php?topic=4370.msg30661#msg30661
Geoff

[dayers]

Hi Robert,

I have discovered that the preview for using BN can be quite small. Finding a large enough patch of "sky" is a challenge, though, for targets like this. For CC, I assume you are using the entire image for white reference. I agree with Geoff that lining up the R, G, and B peaks in HT often helps in bringing out star color. Please keep us posted on how you end up solving this problem.

Dave

[Alejandro Tombolini]

Hi Robert, do not use ABE in this case of extended faint nebula. DBE will work better.
First duplicate the image, use HistogramEqualization on it to have an idea on where place the samples, generate an icon and then apply DBE icon over de original image.
For colours in stars, deconvolve the image and wait until saturation to see if the colour is good.

Regards,
Alejandro. 

[robertm]

Thank you for those tips and suggestions, I'll give them a go.

Thanks
Robert

[robertm]

Well I thought I'd let you know what the solution was even though it wasn't anything to do with Pixinsight.  The last 8 subs had a blue colour cast probably due to the pending dawn even though it looked dark to my eyes.  I just checked the total background counts for each of the channels and saw the difference, mainly in the blue channel after sub 120.  Now these are excluded the colour does balance quite well.

Thanks for your replies.

Robert

[Alejandro Tombolini]

Hi Robert, good to know it was solved.

Regards, Alejandro.