photometric color calibration : stars excluded...?

Software Siril
1

Hello,

I run the color calibration after preprocessing my original images by simply subtracting the master bias.

The console output indicates all is going fine, until it excludes all the stars that it had found and matched with the catalog :

20:14:28: Solving on selected area: 0 0 3008 3008
20:14:28: Findstar: processing…
20:14:29: Catalog NOMAD size: 126 objects
20:14:29: 70 pair matches.
20:14:29: Inliers: 0.986
20:14:29: Resolution: 1.420 arcsec/px
20:14:29: Rotation: +95.90 deg
20:14:29: Focal: 546.09 mm
20:14:29: Pixel size: 3.76 µm
20:14:29: Field of view: 01d 11m 11.97s x 01d 11m 11.97s
20:14:29: Image center: alpha: 15h05m10s, delta: +62°16’11"
20:14:29: Normalizing on red channel.
20:14:29: Applying aperture photometry to 69 stars.
20:14:33: 69 stars excluded from the calculation
20:14:33: No valid stars found.

The histogram shows that there are pixels with negative values. Also the ‘stat’ command shows that the min value is negative :

20:30:37: Running command: stat
20:30:37: Red layer: Mean: 37.4, Median: 35.5, Sigma: 71.5, AvgDev: 11.9, Min: -117.0, Max: 45897.0
20:30:37: Green layer: Mean: 54.9, Median: 51.6, Sigma: 92.4, AvgDev: 13.2, Min: -115.0, Max: 56679.8
20:30:37: Blue layer: Mean: 25.5, Median: 23.8, Sigma: 56.1, AvgDev: 11.1, Min: -117.0, Max: 38328.6

I use the command “offset 200” to make all values positive
When I then run the color calibration again it works perfectly :

20:33:21: Solving on selected area: 0 0 3008 3008
20:33:22: Findstar: processing…
20:33:22: Catalog NOMAD size: 126 objects
20:33:23: 71 pair matches.
20:33:23: Inliers: 0.986
20:33:23: Resolution: 1.420 arcsec/px
20:33:23: Rotation: +95.89 deg
20:33:23: Focal: 546.09 mm
20:33:23: Pixel size: 3.76 µm
20:33:23: Field of view: 01d 11m 11.98s x 01d 11m 11.98s
20:33:23: Image center: alpha: 15h05m10s, delta: +62°16’11"
20:33:23: Normalizing on red channel.
20:33:23: Applying aperture photometry to 70 stars.
20:33:32: 5 stars excluded from the calculation
20:33:32: Color calibration factors:
20:33:32: K0: 1.000
20:33:32: K1: 0.661
20:33:32: K2: 1.087
20:33:32: Background reference:
20:33:32: B0: 3.59287e-03
20:33:32: B1: 2.53904e-03
20:33:33: B2: 3.70967e-03

Does the photometric color calibration not tolerate negative pixel values ?
Or did I make an error in preprocessing ? (My images are 32 bit after the preprocessing)

Many thanks for your help , Jan

2

Hello, in version 1.0, stars that had negative pixels around them were exluded indeed. If you have version 1.1 you can allow it in the preferences.
The median of you image is rather low, I’d say they are either incorrectly calibrated or shot with a very low offset setting. If it’s the latter, you could increase it a little.

3

Hello Vincent,
thank you for confirming my suspicion!
I am currently using version 1.0.6. I think that my calibration is correctly done because it is so simple. My camera ASI533mc is cooled and I do not subtract darks anymore, only bias. My offset is 20 (and gain 100) , which leaves enough margin at the low end of the histogram: the median value of the bias is around 797. The problem comes from the light frames, in which the some pixels have values lower then in the bias frame . Therefor increasing the camera offset will probably not remove the negative values after subtraction. As a workaround, I will subtract a small constant from my master bias such that after preprocessing all values remain positive.
Siril is a fantastic software and I am curious about he new features coming in the next release!
Again many thanks for your support!

4

Make sure your bias and light images are taken with the same offset and gain. Indeed lights have more noise because they have a longer exposure, and if there is no significant thermal current it their values will fluctuate around the offset level. I’m not sure what else you can do about that…