DSLR photometry-is this a sensible approach?

[seeker372011]

I had my first ever go at estimating the magnitude of a couple of bright variable stars this last weekend

I used my modded Canon 300D with a Baader UV/IR filter and took 5 each 30 second images of my targets through my ED80

I also took 5 darks and flats, and pre processed the RAW images and stacked them as I normally would in Deep Sky Stacker.

I then took the image to IRIS and split the channels

Working off the green channel only, I used IRIS to estimate the intensity-ADUs-for the target and 4 to 6 comparison stars (after subtracting background sky intensity)

I used a AAVSO visual chart as I was only testing the approach not trying to collect reportable data just yet.

I then plotted the intensity vs the magnitude for the comparison stars (since this was a visual chart the magnitudes of the comparison stars were only to one decimal place).

the readings appeared reasonably linear, so I fit a best fit line and derived an equation for the line (the R squared was 0.97 and 0.98 respectively)

I then used the measured intensity in ADUs , plugged into the equation derived above to estimate the magnitude of my target star-to one decimal place.

so-is there some fundamental flaw in this approach or is it methodologically sound?

I do not have a mono camera and do not have a V filter. And I did not use a transformational co-efficient-not in the least because I have no idea how to go about calculating one ;)

thanks in advance for any comments or advice. As I said above this is my first attempt at variable star photometry

Narayan

[[1ponders]]

There are a few thing it will depend on Narayan, but primarily at what point does non-linearity start with your camera. If you are interested I have an assignment I did on just this topic. pm me your email address and I can send it to you (about 1.2 mb)

[seeker372011]

PM sent

[Terry B]

Yes you can do it. There are a number of papers about using the green channel as a defacto "V" filter but I'm on holidays and don't have the links with me.
One rule though is it is preferable not to stack images and you must not perform any sort of stretches etc on the image.
You can median stack or average images but it is preferable to measure the flux on the individual images and average the results.
It is very easy to get the "V" fluxes from the AAVSO charts and try some measurements.
Good luck.

[Bassnut]

Well, ironically, I know nothing about Photometry and yet my GRAS rig is largely used by Photometrists because it has a NABG cam. From what I understand, linearity is paramount, they are loathed to use a ABG cam, even if its a premium cooled CCD.

[seeker372011]

woo that's good advice though that just quintiplied-if there is such a word -the effort :(

I did use median stacking of course, and did not do any sort of stretching

would appreciate any links whenever you get the chance..especially about how to go about estimating the transform co-efficient


narayan

[seeker372011]

yes but apparently if you stay far away from where the stars get saturated one should be in the linear region..my results plotted scarily close to a straight line

narayan

[Bassnut]

You may well be right. Apparently tho, ABG cams continuously skim the CCD of stray photons (hence the lower QE) affecting linearity over the whole dynamic range. I might be getting picky here, again, I dont know really, perhaps the ABG function is roughly linear up to saturation, as you report.

[seeker372011]

It has been pointed out to me that it could not have been absolute intensity I plotted against magnitude to get a straight line, and of course this is absolutely true

I plotted -2.5 log (intensity) vs Magnitude. Should have said so in the first place

Narayan

[Terry B]

Here is the link I mentioned.
There are others also.


http://www.socastrosci.org/2007%20papers/Hoot_DSLRPhotometry.pdf

[Merlin66]

Here's the AAVSO link on CCD photometry, if you haven't already found it!
http://www.aavso.org/observing/programs/ccd/manual/index.shtml#new

You might want to invest in a Schuler V filter from AstroDon, I think around $60 or so.

[seeker372011]

thanks..the bottom line appears to be that DSLRs are OK for differential photometry and at a pinch could be used for survey as well..which is reassuring

[seeker372011]

unfortunately Astrodon seem to have doubled the price-used to be $75 or so but now it is $150 for the 1.25 inch

Ouch

so I am looking on astromart for a used filter if I can get one

[Merlin66]

Try Astromart, I picked up a full set (5 filters) for only $200 just before Xmas.

[citivolus]

I would think that for maximum accuracy you would want to extract the green channel data pre-debayering, so that it is not including interpolated results. Having not done this myself, I do not know what kind of impact this would have on measured luminosity.

[seeker372011]

I am not sure how I could do this-extract the green channel prior to debayering? I have IRIS and Nebulosity as my software packages?

[Terry B]

With iris it is very easy.
2 choices:
1. when you select the "decode raw file" option just press the "-->RGB" button and the 3 channels are created. or
2. use the RGB separation option on the already decoded file.

[seeker372011]

just to clarify:

I am using 2-which is after de-bayering..so if I use method 1, it will split the channel without debayering?

thanks, sorry to be dense!

[Terry B]

Debayering and splitting the channels are essentially the same thing.All you are doing is to select out the pixels that have a green filter in front of them.
I just had a play in iris and it doesn't seem to matter what order you do it.
What you mustn't do is the option of making a B&W image from the original raw file. This uses an algorithm that used all 3 filters and is not the same as just the green channel.

[seeker372011]

..yes of course it makes sense

many thanks

taken my next set of observations now..this variable star observing can be quite addictive

[higginsdj]

Some points to bear in mind:

1. Do a linearity test BEFORE doing anything else. Do this for the channel that you will be measuring (try it for R and B as well). R may be better. AGB cameras do have some linearity. It just a matter of finding out what that 'exposure' band is.

2. I assume your UV/IR filter is a blocker in those bands rather than allow only those bands.

3. There is nothing wrong with stacking images for photometric use, just make sure you Dark, Flat and Bias correct them first.

4. G is only very loosly like V. G is a much broader band with 'leaks' than V. In fact you way wish to convert your RGB layers into mono and stack them as a single mono to replicate a "Clear" filter to give you better signal strength.

5. Precision is about SNR and linearity. If you are not calibrating with flats, darks and bias then can only achieve +/- 0.5 mag at best (no matter what your curves say) due to low SNR.

6. Your on the right track and there is nothing stopping you doing precision differential photometry with a DSLR if you know your systems behaviour and stay within it's boundaries.

If you want addictive, start working an eclipsing binary star and plotting your results. You'll get a usable period plot after observing it continuously for 4-8hrs.

Cheers