Image integration question with very low noise.

[zackyd]

Hi everyone, I just started learning PI and CCDs and have a few questions. While it is cloudy I am trying to understand how PI works with the Dark frames and bias I have been capturing. I am currently trying to get a "black" frame by using my individual dark frames and integrating them together after calibration. with the master bias and dark. I am failing to produce a completely noise free image on the last integration of the calibrated darks. I have something like 17 frames that have very low amounts of random thermal pixels and cosmic rays, and I am trying to integrate them to see how much of a noise free image I can produce.

I have started with pretty much the default values in the integration and tried quite a few combinations and have not been able to reject the hot pixels and cosmic rays.

Which settings will I need to focus on?

Thanks for any help and my apologies for a somewhat strange request

[RickS]

With 17 images Winsorized Sigma Clipping would be a good choice as the rejection algorithm.  Then you'll need to adjust the Sigma High and Sigma Low sliders to vary the strength of the rejection.

It might be worth having a play with CosmeticCorrection as well.  It deals with the pixels and columns that don't behave predictably enough to calibrate out completely.

How are you judging the quality of the result?  Even low amounts of noise will look ugly if you stretch the data hard enough.

Cheers,
Rick.

[zackyd]

This are extreme low noise as far as I can tell when stretched. I am still

here is a link to download a few of the frames if you would like to get an idea of what im looking at.
http://we.tl/MQGLYZlnzw

I have moved the high and low, and cant for the life of me get it to clip out the random noise upon integration.

Thanks,
Z

[RickS]

I have moved the high and low, and cant for the life of me get it to clip out the random noise upon integration.

I think you might be setting your expectations a bit high if you hope to remove the random noise with a rejection algorithm.  Rejection will deal with unwanted signal, e.g. cosmic ray hits, but it won't remove low level random noise.  To reduce noise you need more data.

Since you're playing with dark frames your data is inherently undithered.  That means that hot and cold pixels that aren't removed by calibration probably can't be cleaned up by rejection either.  With dithered light frames you will get better results.

Cheers,
Rick.