Registar and CCD hot pixels/column defects

[jase]

I've been using Registar (http://www.aurigaimaging.com/) for about two years on different data - both CCD and film scans. Very powerful and flexible image registration (alignment) tool. However I have noticed of late that when using the software to align multiple CCD images, I'm having a few problems with hot pixels and column defects.

Typically, I dither each exposure by 3 or 5 pixels (this is performed automatically by MaximDL and the autoguider) so that when I median combine the images, such defects appear randomly across the frame. As they are random, the median combine routine removes them - sounds simple and logical right? However, I suspect that because Registar's registration process is so accurate, there are situations where column defects and hot pixels occur in the same location in each frame. Thus the median combine considers the defect normal and includes them in the combined final image.:( Sure, it sounds great to have accurately aligned sub exposures, but all defects are also aligned:shrug:. I have not tried a SD mask or sigma reject combine process, but I'd imagine this would be similar.

This isn't a major issue as I can always use the healing tools in PS, but this takes time that could otherwise be used elsewhere. Ultimately, this is the reason why I dither each exposure to counter act pixel/column irregularities in the first place! I guess I could produce a bad pixel map, which may partly address the situation.

Anyway, I thought I'd ask if anyone here has overcome similar issues before I go to the source. Perhaps I need to increase my dithering size, but if Registar is going to realign the sub exposures based on thousands of stars with absolute perfection, I can't see much benefit.

:help:

[gbeal]

A bit more technical than necessary for me at the moment, but I did wind back the dithering, as between sequences it caused major trauma with the auto-giuding, in my case. In the end I was faced with either no dithering (which I elected to go) or a delay between sequences of something like 60 seconds. My "usual" darks or subs have a decent column defect, which is apparently quite normal, and I haven't noticed anything untoward, but then again I haven't used Registar either.

[jase]

Thanks for the feedback Gary. I don't have any issues with dithering as such. My guider pics up the guidestar between sub exposures rather quickly and makes adequate adjustments to return to "normal" guiding levels even with 5 second exposure times.

I've pinged Registar tech support an email. Not that its a fault with their product. Its doing its job (too well in this case), but they may provide something to work on. I may post on some other forums if I don't get any traction.

Thanks again.

[jase]

I got a response back from Auriga Imaging. They indicate that if the sub exposures have the column defects/hot pixels in different areas of the image (achieved by pixel dithering between subs), they will still be successfully removed on a median (or sigma reject) combine function. So the image registration as such will not impact this. Two additional comments were made, take more subs and check dithering configuration.

I think on reflection, the latter has got me this time around. When I set up a imaging sequence in MaximDL, it keeps track of the dithering used. If you stop a sequence (which I did due to clouds), then start it again, dithering zeros (defaults) and starts from the same location again. Thus not all sub exposure defects were in different locations. The more subs you take the better this will be handled. I may hit some MaximDL forums to valid the dithering process.