One shot color flats

[bpipes]

I am trying to move from Maxim DL and Photoshop (in Windows) to PI (in Mac OS) and have run into something I cannot figure out. Ron Wodaski developed a plug-in for Maxim DL called Color Flat Maker that normalized raw flats from OSC's so they could be used on Bayer pattern images. How should I handle OSC flats in PI?

Bruce Pipes

[bpipes]

Well, the silence is deafening. Ron Wodaski's website says that the source code for his color flat maker plug-in is free. I wonder if someone who knows how could turn that into a script for PI. Just a thought.

Bruce

[Nocturnal]

Bruce,

have you tried the stacker script that someone developed? Perhaps it can handle OSC flats. I use DeepSkyStacker which does all that magic automatically and I'm not sure how to do the same in PI, sorry.

[pfile]

well, i use a DSLR, which i guess qualifies as an OSC, and i've never heard of having to normalize the flats beyond what is done during the stacking of the flat subs.

that does not mean i know what i am doing, though.

[bpipes]

Sander and pfile,

Thanks for the replies. The idea behind normalizing OSC flats is to deal with the potentially widely varying brightnesses of the red, green and blue pixels in the flat frames. Some folks blur their flats for OSC images, and that is better than nothing.

Give the sophistication of PI, I presume at some point a color flat normalizing process will be added. I guess for now I will normalize my flats in Maxim DL.

Best regards,
Bruce

[Nocturnal]

Hi Bruce,

this is how DSS deals with OSC flats. It normalizes each channel separately which results in a raw (undebayered) master flat. This is also how it's described in the handbook of astronomical image processing.

I don't know how PI builds a master flat out of raw OSC frames and how it's used to calibrate, sorry.

[Juan Conejero]

Hi Bruce,

Sorry for the delay in my answer —I just can't keep up to date with the forum, due to my current workload.

Just select your master flat frame (undebayered) in ImageCalibration; that's all. You need no normalization scripts with PI. The flat frame will be normalized by internal IC routines, and any residual color casts can be easily corrected after calibration and integration with the BackgroundNeutralization and ColorCalibration tools.

[bpipes]

Hi Bruce,

Sorry for the delay in my answer —I just can't keep up to date with the forum, due to my current workload.

Just select your master flat frame (undebayered) in ImageCalibration; that's all. You need no normalization scripts with PI. The flat frame will be normalized by internal IC routines, and any residual color casts can be easily corrected after calibration and integration with the BackgroundNeutralization and ColorCalibration tools.

Thanks. Juan. That is what I wanted to know. I am getting severe color casts after calibration and integration, and I thought that might be it. I'll keep working on it.

Bruce

[oldwexi]

Hi Bruce!
I do have also bad background color results with my Sky-Flats shot after sundown.
They are to blue. And as a result despite all promises i get calibrated images where the blue channel
is much to dark (with a blue data loss) in the calibrated and stacked images. I built a workaround for that.
So, to get single flats without a blue color cast i simply move the pixels of a flat with 0.5 pixel in x direction and y direction.
This averages the blue, green and red RAW pixels. It gets me a flat with 1 histogram bump and not with 3 histogram bumps.
And,  i get a nice flat without blue color cast but the dust speck still are there precisely.
The procedure i do is the following in detail:

Step 1 use RAW Flat Image    Process  „ConvertToRGBcolor“  (Produces a Pseudo(!) color image)

Step 2 use result of Step1     Process “Channel Match” 0.5x 0.5y (Shifts the pixels 0.5 pixel in x and y)

Step 3 use result of  Step2    Process “ConvertToGrayscale”  (Returns the (Pseudo)color image back to the gray OSC RAW Flat image)


Using these flats gets me much better color results after calibration and stacking. The background has very similar values in all colors.
Here the Link to the XPSM file with the above steps for simple drag and drop in PI.
http://www.werbeagentur.org/oldwexi/Daten/Corr_Flat_Field.xpsm

Gerald

[bpipes]

Thanks, Gerald. I'll give it a try. Wodaski's color flat normalization does just what you say, i.e. converts a raw flat with three humps to a normalized one with one hump.

Bruce

[Carlos Milovic]

Hi Bruce
I'm using a Canon DSLR, and I never worried about the different scalations. ImageCalibration sorts it quite well.