Hello all,
I have a standard unmodded Canon 450D which I now use for imaging VSO and other objects, without any filters, flats, etc. I would like to do "quick and dirty" magnitude estimates directly from these color images.
I would just inspect the size/brightness of the target on the images using suitable comp stars, brighter and fainter, similar to how I do visual observations. My question to the forum, is this a good idea?
I can see that my error by this method will likely be 0.15 - 0.20 mags vs. 0.05 - 0.10 by direct visual, due to uncertainties in the pixelation and image quality, etc. vs direct visual is pretty accurate.
1. How should I report it in WebObs, given the "VisDig" option is not available? Some prior suggestions were do it as a CCD observation with "Unfiltered V zero point". Is this the best way?
2. Given lower accuracy of this method vs. visual, is it generally allowed, or not particularly recommended from AAVSO management point of view?
Thanks for any and all input!
Mike
Hi Mike,
I do this occasionally. Prior to doing this I make sure none of the stars I am estimating in my image are not saturated. I report it as a visual observation.
Barbara
Interesting thought
.
I just checked and "Visual from Digital Image" is (now?) available in the WebObs drop-down box to select the observation type.
As they are color images, you are using filters, actually three of them (Tricolor R,G,B). Since the green filter of DSLRs are really close to Johnson V, you should get better results by extraxcting just the raw- green channel, and then use this with reference magnitudes in Johnson V, and report the result as "Tricolor G", so actually use a similar workflow like people doing aperture photometry with DSLRs.
This means that to make the best use of the material from your DSLR, you should use the RAW (*.CR2) files that the camera can produce if you tell it to also store the raw images (not just JPEGs), and then use some software that can load the RAW and can separate the three color channels, like for example FITSWORK. So:
1) Load RAW *.cr2 image in fitswork
2) "Debayer": create a color image from the raw image without applying an extra white balance to preserve the raw nature of the image: Processing => Bayer Interpolation => Demosaic raw image (in the dialog deselect "Camera Colorprofile" and Enter 1.0 for Rmul, Bmul, G1mul (neutral whitebalance). Pressing "Generate Image" will create a color image. Don't worry if thr image has e.g. a greenish color bias instead of a black sky background, that doesn't matter.
3) extract the three colors , e.g. in fitswork use the menu Processing -> Split color Image
You get three images, probably the useful one is the green one. Perhaps it's easier to extract the magnitudes visually if you invert the image (as if looking at photo negatives).
Clear Skies
HBE
The issue of VISDIG is a bit nuanced. It exists as a category, but is rarely used. It broke at some point years ago, and we have been unable to put the manpower towards fixing it. As such, I agree with Barbara that the best idea is to submit these as visual. However, in the case of DSLR this is much more complicated because the human eye is a filter and I have no idea how close the color DSLR image is to that. So, if you wanted to submit a DSLR image in particular, the best idea would be to do as HBE suggests. This is now no longer simple and requires a lot more effort. It would be worth comparing the magnitudes you get to Visual magnitudes and seeing how close you come. If it appears to be largely similar, then you can use it and the filter differences are just incorporated in the error. However, if they are significantly different, then it submitting them is likely not recommended. I guess what I would recommend is to proceed with caution and see how things go. Hope this helps.
Thanks,
Bert Pablo
Staff Astronomer, AAVSO