Star Alignment

[sreilly]

I had included 8 images of M51 to align and used a previously aligned image as the reference. On first attempt the alignment failed on each image. Second attempts, same series and same process instance, it all worked.

Below is a segment of the Process Console:

Registering target image 1 of 8
Loading target file:
G:/Images/M51/PI_Cal/M51-001-L-090-E_dupe-12_c.fts
Reading FITS image: 32-bit floating point, 1 channel(s), 4008x2672 pixels: done
153 FITS keywords extracted
Structure map: done
Detecting stars: done
494 stars found.
Matching stars: done
*** 0 star pair matches found - need at least six matched stars.
* Previous attempt failed - this is try #2
useScaleDifferences=false
Matching stars: done
295 putative star pair matches.
Performing RANSAC: done
279 star pair matches in 220 RANSAC iterations.
Summary of model properties:
Inliers     : 0.946
Overlapping : 1.000
Regularity  : 0.974
Quality     : 0.886
Root mean square error:
delta_RMS   :  0.755 px
RMS error deviation:
sigma_RMS   :  0.287 px
Peak errors:
delta_x_max :  1.935 px
delta_y_max :  1.567 px
Transformation matrix:
     +0.9743     +0.0024    +54.2292
     -0.0028     +0.9743    +52.9173
     +0.0000     -0.0000     +1.0000
scale       : 1.027
scale-X     : 1.026
scale-Y     : 1.026
rotation    :     -0.15 deg
dx          :    +54.23 px
dy          :    +52.92 px
Generating registered image
Homographic Projection / Lanczos-3 interpolation, c=0.15: done
Registration successful.
Writing output file: G:/Images/M51/PI_Cal/aligned/M51-001-L-090-E_dupe-12_c_r.fit
174 FITS keywords embedded
Writing FITS image: 32-bit floating point, 1 channel(s), 4008x2672 pixels: done

Registering target image 2 of 8
Loading target file:
G:/Images/M51/PI_Cal/M51-001-L-090-E_dupe-13_c.fts
Reading FITS image: 32-bit floating point, 1 channel(s), 4008x2672 pixels: done
153 FITS keywords extracted
Structure map: done
Detecting stars: done
460 stars found.
Matching stars: done
*** 0 star pair matches found - need at least six matched stars.
* Previous attempt failed - this is try #2
useScaleDifferences=false
Matching stars: done
290 putative star pair matches.
Performing RANSAC: done
273 star pair matches in 166 RANSAC iterations.
Summary of model properties:
Inliers     : 0.941
Overlapping : 1.000
Regularity  : 0.974
Quality     : 0.882
Root mean square error:
delta_RMS   :  0.775 px
RMS error deviation:
sigma_RMS   :  0.310 px
Peak errors:
delta_x_max :  1.974 px
delta_y_max :  1.543 px
Transformation matrix:
     +0.9740     +0.0030    +72.0915
     -0.0033     +0.9744    +42.8474
     -0.0000     -0.0000     +1.0000
scale       : 1.027
scale-X     : 1.027
scale-Y     : 1.026
rotation    :     -0.17 deg
dx          :    +72.09 px
dy          :    +42.85 px
Generating registered image
Homographic Projection / Lanczos-3 interpolation, c=0.15: done
Registration successful.
Writing output file: G:/Images/M51/PI_Cal/aligned/M51-001-L-090-E_dupe-13_c_r.fit
174 FITS keywords embedded
Writing FITS image: 32-bit floating point, 1 channel(s), 4008x2672 pixels: done

[sreilly]

Seems to have happened after latest update. Running version PixInsight Core 01.08.04.1195 Ripley (x64) on Windows 10 64 Pro Screen shot is below:

[Juan Conejero]

This is just normal behavior. Somehow you had SA's use scale differences parameter enabled. Since there are scale differences between the images being registered and the reference image (about a 3%), registration fails with this parameter enabled. SA's automatic probing feature has performed a second attempt with the same parameter disabled, which has worked perfectly. As noted, this is the normal and expected behavior, so there is no bug.

[sreilly]

Thanks Juan,

I have no idea why that would have changed as I have always used the default settings. I noticed this was under Mosaic Mode Only. As far as I know this has always been the settings I used and never got the failed first attempt. But as long as this is expected behaviour we should be good. I have noticed with the last few updates this past week PI has started and worked faster as well. Thanks for all the hard work in keeping this the #1 processing software. Sorry to see there weren't any PI workshops at NEAIC/NEAF this year.

Steve

[Juan Conejero]

Thanks for all the hard work in keeping this the #1 processing software.

Thank you for your support Steve!

Sorry to see there weren't any PI workshops at NEAIC/NEAF this year.

... but not because we are unwilling.