Linear Fit Clipping vs GESD pixel rejection - Bias Frames -- % Rejects

rockenrog

Member
Apr 7, 2018
20
0
Your experienced advice/comments appreciated...
In Sept 2020 I am re-processing a 2019 image, now learning the Weighted Batch Preprocessing. Pixel rejections are quite different vs last year.....
I created a master bias in Sept 2019 using 103 bias frames using Image Integration & Linear Fit Clipping setting of Lfit High = 2.5 and Lfit Low = 5.0.
PI 1.08.06.1457.
Result is 1.1% High rejects and .04% Low rejects.

I just used the same 103 images and pixel rejection using GESD (with standard parameters and via WBPP v1.4.8).
PI 1.8.8-6.
Result is 5.8% High rejects and .003% Low.

Does it make sense that bias frames could have this high percentage of outliers?

I am using QHY168C CMOS camera at -10C.

I am still checking the Darks and Lights and not ready to comment about them.

Thanks,
Roger
 

ngc1535

PTeam Member
Feb 1, 2014
559
77
AdamBlockStudios.com
Roger,

If I understand your question, you are trying to compare the "thresholds" of these two algorithms? They certainly will not give you the same answer. If you used the same algorithm *and then* got two different answers for the same thresholds, that would be something to investigate. What I am trying to say is that the different percentage is due to the algorithms (and their effectively different thresholds on what qualifies as an outlier) and not something about your data.
I could be misunderstanding the question though.
-adam
 

rockenrog

Member
Apr 7, 2018
20
0
Roger,

If I understand your question, you are trying to compare the "thresholds" of these two algorithms? They certainly will not give you the same answer. If you used the same algorithm *and then* got two different answers for the same thresholds, that would be something to investigate. What I am trying to say is that the different percentage is due to the algorithms (and their effectively different thresholds on what qualifies as an outlier) and not something about your data.
I could be misunderstanding the question though.
-adam
Adam,
Yes, I am comparing the thresholds for two different algorithms. I was surprised that with uniform data bias frames, that GESD would result in 5% outliers. From your reply, it seems that my result is not unexpected.
So I will continue my learning with Weighted Batch Preprocessing.
My study is using the raw data that I manually processed last year (to learn how to do it properly) with good results. I hope I can get WBPP to do as good a job. Let' see.
Roger