Page 1 of 1
Preprocessing artifacts
Posted: 17 Dec 2018, 15:28
by Forum_2018
i've noticed repeatedly after preprocessing the resultant image looks like the pixel arrangement is always diagonal from bottom left to top right - see attachment
Is this a result of the stacking algorithm and should I be worried?
Re: Preprocessing artifacts
Posted: 17 Dec 2018, 15:29
by Forum_2018
No one...not even Fabio?
Re: Preprocessing artifacts
Posted: 17 Dec 2018, 15:29
by Forum_2018
Hello, I missed the message.
The cause seems to me an imperfect dark frame correction, which gets a diagonal shape cause the auto-alignment of several images. If you orient the camera with the equatorial axes the noise should get a horizontal shape.
Remember that the master dark frame must be at least the average of 4 frames (8 or more is better).
If you are using a bias as a dark frame, then again you should average several frames.
The image also seems extremely oversampled: you may work in binning 2x2 without losing any information.
Greetings,
Fabio.
Re: Preprocessing artifacts
Posted: 17 Dec 2018, 15:30
by Forum_2018
Here is a photo from a book about astro photography by Ron Wodaski that shows similar artifacts due to stacking very short sub exposures!
Iver
Re: Preprocessing artifacts
Posted: 17 Dec 2018, 15:30
by Forum_2018
Thanks Fabio. I was using multiple dark frames. What do you mean by over-sampled?
Re: Preprocessing artifacts
Posted: 17 Dec 2018, 15:30
by Forum_2018
thanks Iver.
Re: Preprocessing artifacts
Posted: 17 Dec 2018, 15:31
by Forum_2018
An image if oversampled if it contains "too many pixels" (or too high resolution) for the information that it really contains, see:
https://www.cloudynights.com/topic/3009 ... -an-image/
A test:
1) Open (in any graphic software) a good image.
2) Resize it to 50%.
3) Now resize it to 200%.
the result will have the same size, but it will be a blurred image, because step (2) decreased the information of the image.
However, if at the step (1) you start with an oversampled image (e.g. blurred from the beginning) then after those steps, the result is as good as the original.
This means that an oversampled image can be halved in size safely (e.g. work in binning 2x2) and processed at this size, without losing any information. Then, if needed it can be doubled in size as a last step.
Working on oversampled images is a waste of processing time and/or disk space.
This is the same thing of "megapixel myth" (Google it) that affected consumer cameras in the last years.
Re: Preprocessing artifacts
Posted: 17 Dec 2018, 15:31
by Forum_2018
It is commonly called "walking noise" Try dithering your exposures and make sure you are well polar aligned.
John
CCD-Freak