Debayering a DSLR's Bayer matrix.

[alistairsam]

i mustve gone over the blue area, picture is completely black,

oh well, back to square one

[Gina]

Got the cover glass off but not quite in one piece - a corner cracked off.  Having got over half the bond broken I put the filter frame back complete with the very sticky double sided tape.  I pressed it all together firmly by hand and then took the craft knife and carefully separated it again.  This time the cover glass stuck more firmly to the double sided adhesive tape than the sensor frame - et viola the sensor is clear

[Gina]

You can insert the quote tags manually and it appears to handle things properly.  It's a bit of a pain though

James

I tried that though I may not have got it right...

[Gina]

i mustve gone over the blue area, picture is completely black,

oh well, back to square one

Oh dear - sorry to read that Alistair   I know exactly how you feel!  AARRGGHH!!

[Gina]

I've just won me a 450D    I really must stop this reckless spending spree!!  I ought to save my money for my EQ8 or Esprit 80ED or Borg 77ED or ...

[Gina]

Bebayering 1000D sensor is going well

 

[Luis Campos]

i mustve gone over the blue area, picture is completely black,

oh well, back to square one

I'm really sorry to hear that buddy...

[Luis Campos]

Bebayering 1000D sensor is going well

CFA Removal 18.jpg  CFA Removal 19.jpg

...Practice leads to perfection, right Gina?

[Gina]

Got some images with the mono 1100D on the MN190 last night as it was reasonably clear for a while.  Again with the Baader 7nm Ha filter.  Focussed up on the moon and got some shots at the same time then slewed to the Eagle Nebula and took a few at different exposures.  The one below is 10m at ISO 6400.

1. Moon cropped just to remove the CFA border
2. Cropped more
3. Actual image size pixel for pixel
4. ?? Space junk?  Moon Lander?  Or sensor damage?
5. Eagle Nebula

       

No. 4 zoomed in x4.

Edited September 16, 2013 by Gina

[Gina]

It's sensor damage.  Moon Lander ruled out because it's in a different place on the moon in another moon image.  It's too sharp to be outside the camera too.

[Gina]

Here's a flat cropped to cut off the part where the CFA remains.  Stretched to show the variations in sensitivity.

[foundaplanet]

It's sensor damage.  Moon Lander ruled out because it's in a different place on the moon in another moon image.  It's too sharp to be outside the camera too.

Best hope the conspiracy theorists don't get hold of it.!

Ed: Though it does look a bit like an oil rig casting a shadow..

Edited September 16, 2013 by foundaplanet

[Gina]

More 1000D CFA removal - still working

1. Light out of the window
2. Flat as taken
3. Flat stretched until the debayered zone stretched from black to white

   

Edited September 16, 2013 by Gina

[Gina]

And a 2 minute dark at ISO 1600 (max for the 1000D) stretched enormously in Ps until the noise showed.  A slight hint of increased background signal here and there but nothing significant and no added hot pixels.

Edited September 16, 2013 by Gina

[Luis Campos]

And a 2 minute dark at ISO 1600 (max for the 1000D) stretched enormously in Ps until the noise showed.  A slight hint of increased background signal here and there but nothing significant and no added hot pixels.

CFA Removal 23.jpg

This one looks far better then the other attempt, much more linear no doubt, I think this one will work great for you

Congrats Gina, it sure is a well deserved victory

Cheers,

Luís

[Gina]

Thank you Luis   I think I'll try cooling with the 1000D though I'm aware that it isn't as good as a 1100D or 450D.  I should have a 450D coming tomorrow - I'll be interested in seeing what differences there are in construction from the 1100D.  The 1000D seems very similar though may need a slightly deeper dry chamber if I use the same design as the 1100D.

[Gina]

More 1000D CFA removal - I don't think I dare go any further.  Not unless I can find a way of making a finer CFA removal tool.  I think I might have caught some of the blue border or something because I seem to have attracted a dead column and a suspect row.  Otherwise the results are still not bad and I haven't managed to kill this one   I'll do a dark tonight as the camera body is not light tight enough for ultra long exposures in daylight.

1. View out of the window slightly improved in Ps
2. Flat stretched "till it squeaked" to bring out the defects.

 

[Luis Campos]

Oh no....

I hope it doens't affect long exposure perofrmance, the same happened to one of my 350D's sensor, one too deep scratch and I got two dead pixels collumns, it still works but long exposure is seriously compromised, it shows horrible noise...

Best of luck Gina

Cheers, 

[Gina]

Thanks Luis   I tried to take some darks last night but kept getting Error 99 which I think is to do with the lens connection - not sure.  Anyway, I'll put the sensor in another body and see what happens.

[Gina]

Working fine in the other body but I can't take proper darks until erm... dark.

[Gina]

The 450D came this morning so I'll be taking a look at that.  I'm wondering though about using the 450D for my general photography instead of an 1100D which I'm currently using as I already have 1100D parts carved up for cooling.  Then using the 1100D sensor for debayering and cooling.  I'll take the 450D apart though anyway and that may determine things.

Is the 450D sensor easier to debayer than the 1100D?  I seem to remember reading that it was but without reading through over a thousand posts I'm not sure.

Edited September 18, 2013 by Gina

[Gina]

Been looking at the 450D spec.  Comparing it with the 1100D.  Max ISO only 1600 as compared with 6400.  Earlier image processor.  Otherwise much the same.  Had a look at the insides and it looks very similar to the 1000D.  So much so that I had to check that this is a 450D - actually a Kiss X2.  For astro use the 1100D wins.  When I was using the cooled 1100D in earnest (before buying Atik CCDs) I used ISO 3200 mainly with 6400 for framing and focussing.  Exposures up to an hour at 3200 were possible.  Of course, compared with the Atiks the long exposures were necessary but with debayering I thing the difference may be reduced quite considerably.

[Gina]

I put all the screws back into the case of the 1000D and covered the viewfinder with the supplied rubber cover.  That has stopped the daylight getting in and I've taken a 2m dark at ISO 1600, converted it to TIFF and stretched it enormously in Ps.  I'll do a longer exposure shortly.

[Gina]

Here's a 5 minute dark stretched in Ps.  This is taken with just the camera as having the USB plugged in lets light in.  Later on, when it's dark I'll take some longer ones and see how that goes.  I might also take some lights with long exposures.

Edited September 18, 2013 by Gina

[Gina]

I have a couple more sensors on their way (plus non-working bits of camera) a 350D and a 1000D.  Have to say the 1000D looks lije it could be a good option - curreny debayered sensor is showing good results.  Looks good for cooling too