How do you make bad pixel maps for a DSLR?

[almcl]

I've noticed quite a few references to bad pixel maps lately but can't find much, if anything, about how they are created.  

I can see quite a few bad pixels on my subs, but would like to know how to create a map for a Canon DSLR and how they are then applied. I am using DSS to stack then Startools for image processing.

Anyone care to enlighten me? 

[Dave S]

THe bad pixel map is generated by the camera and stays onboard.  Most post proc software deals with hot pixels and you don't necessarily need to do it.

Saying that it won't mess up the camera if it's done (fortunately)

THe bad pixel map is triggered as a function of manual sensor cleaning.

Basically.  Put on the lens cap or body cap and in the menu select manual sensor cleaning.  Let this run for 30 secs then stop the process and restart the camera. Thats's it.

The sensor clean re-triggers the bad pixel mapping and retains it until the firmware is updated when you may need to do it again.

Nothing needs to be exported to processing software.

Dave...

[almcl]

Thanks for that, Dave.

I should have mentioned that the camera is one of Juan's astro-modified ones and no longer has the LP1 anti aliasing filter which has been removed and as consequence Juan advises not using the auto clean function.  However there are two other settings on the clean menu, 'Clean now' and 'Clean manually'.  From what you say, I could have a go with these?

When I've plucked up the courage to try, I'll report back.

[ollypenrice]

My understading is CCD based and derived from the instructions given by the very knowledgeable authors of the AstroArt propgrame. You make a master dark and black clip 2000 ADU from that. This will remove the bias pedestal and noise and leave you with genuine hot pixels. I don't doubt that the precise value of 2000 ADU is negotiable and camera-specific but 2000 works for me. How important a bad pixel map is, I don't know. I use one but I think that the subtraction of a master bias as a dark, along with dither and an aggressive hot pixel filter while stacking using sigma rejection, is really what 'does the business.' I used to use normal darks, but 'bias-as-dark/bad pixel map/hot pixel filtration' beats the classical method hands down for me.

Olly

[Dave S]

6 hours ago, almcl said:

Thanks for that, Dave.

I should have mentioned that the camera is one of Juan's astro-modified ones and no longer has the LP1 anti aliasing filter which has been removed and as consequence Juan advises not using the auto clean function.  However there are two other settings on the clean menu, 'Clean now' and 'Clean manually'.  From what you say, I could have a go with these?

When I've plucked up the courage to try, I'll report back.

No problem. Juan's astro is correct so we don't want to let the sensor clean itself.  Infact no sensor cleaning is involved anyway. You need to select the clean manually function and NOT the clean now option which is another way of saying to the camera go for it and do your cleaning stuff.  Clean manually tells the camera to lock the mirror up so you can poke around on the sensor with puffers and cleaning sticks.  We aren't going to do anything daft like that so it's lens/body caps on to stop any ingress of dust and select clean manually.  We don't want to touch the sensor we just want the camera to build a new bad pixel map.

Dave...

[almcl]

Thanks both Olly and Dave.

I am afraid my ability with DSS isn't quite up to clipping master darks, so that may have to wait a while - although I do dither and am using bias frames (which have the occasional really hot pixel). I may investigate regim as an alternative to DSS and see if that allows a bit more flexibility...

I hadn't realised that the manual clean function also rebuilt the camera's bp map, so here's hoping that's really going to change things!

[wxsatuser]

If your dithering and taking enough subs forget the bad pixels.

Take at least 20 or more subs and sigma clip them in DSS or whatever stacker you use.
This will remove all the outliers such as colour mottle, bad pixels, aircraft trails and even thin cloud.