HDRComposition will not continue or reports bad linear fit.

[astropixel]

HDRComposition aborting process - or reports bad linear fit (lunar images) - output below. No problems otherwise.

Tried various parameters - no change. The script HDRComposition responds the same. Obviously, it doesn't like the images - DSLR_RAW, bias calibrated and debayered. Does the same with the raw cr2 images regardless of the dcraw settings in Format Explorer.

* Integrating HDR image component 2 of 10
Reading FITS: 32-bit floating point, 3 channel(s), 3906x2602 pixels: 100%
Building HDR composition mask: 100%
Fitting linear HDR image: done.
Linear fit functions:
y0 = +0.001178 + 0.883969·x0
y1 = +0.000447 + 0.898127·x1
y2 = +0.000641 + 0.876338·x2
Performing HDR composition: 100%

* Integrating HDR image component 3 of 10
Reading FITS: 32-bit floating point, 3 channel(s), 3906x2602 pixels: 100%
Building HDR composition mask: 100%

** This HDR composition cannot integrate further images - aborting process.

* Dynamic range estimates:
q0 = 2.33156e+05 (17.83 bits)
q1 = 2.64400e+05 (18.01 bits)
q2 = 2.57508e+05 (17.97 bits)
109.556 s

[Nocturnal]

I think maybe they need to be flattened?

[Juan Conejero]

When HDRC fails to build a composite image that's because either:

(1) The images are not aligned correctly, which causes LinearFit (which is part of HDRC) to fail.

or:

(2) One of the images cannot add more data to the "HDR pyramid". In other words, one of the images does not provide low-exposure data to cover previously combined long-exposure data. For example, this happens if you try to build an HDR composition with images acquired with the same exposure times and camera sensitivity.

Assuming that (1) doesn't apply, and assuming that (2) doesn't happen because you're trying to combine a set of similarly exposed images, then you probably need to play a little with the binarizing threshold parameter. Try reducing it.

Edit: Actually, HDRComposition is giving you a clue of what's happening:

y0 = +0.001178 + 0.883969·x0
y1 = +0.000447 + 0.898127·x1
y2 = +0.000641 + 0.876338·x2

These linear fitting coefficients are close to 0.9. This means that the second image in the set is very similar to the first one (only a 10% of the second image is being used in the final HDR composite image). Then the third image will probably have coefficients very close to one, which triggers condition (2) above.

[astropixel]

Thanks Juan - much appreciated. I guessed it had something to do with exposure. They are quite similar. Lesson learned - next time I'll spread the EV more so.

Hi Sander. Should I let that one through to the catcher. What flats?