Should NSL do aperture photometry?
Posted: Sun Oct 24, 2004 3:16 am
I have been wondering whether the NSL project should do "aperture photometry" in addition to the "top pixel count" photometry (e.g. "C9") already computed and posted in the photometry files. Aperture photometry takes an image and has software draw a circle concentic to the center of the source, usually at some factor of the radius of the point spread function (PSF), and find all of the signal inside that circle. Actually, aperture photometry also usually includes more distant circles from which the background would be estimated, but NSL software (WOLF) estimates the background for the entire field all at the same time.
Historically, most photometry is aperture photometry. The reason NSL did NOT use aperture photometry at first was that the PSFs of the CONCAM2s was just so wide that software apertures could usually not be found that there free of the other stars and/or a sloping background.
The CONCAM3s have much smaller PSFs, though, likely making aperture photometry doable. Here is a good link from the American Association of Variable Star Observers (AAVSO) that has a good discussion of CCD aperture photometry: http://www.aavso.org/observing/programs ... al/4.shtml
But should NSL do aperture photometry? What value would it have for us? On the postive side it would likely encorprate more counts than even C25 for many stars, lowering statistical noise, and even pixel noise that bothers both C1 and C5. On the negative side, the extra outlying pixels likely have more error -- both statistical and systematic -- making it possibly less accurate than C25 or C16.
One possibility is that aperture photometry would better allow NSL data to track stars as they rise and set, which could mean the more precise use of comparison stars and the abilty to correct observed counts by a theoretical atmospheric absortion corrrection. I remain skeptical, though, that aperature photometry would really make a big difference here. It will never be able to compete with the accuracy of differential photometry between observations at the same sidereal time on different nights, in my estimation.
- RJN
Historically, most photometry is aperture photometry. The reason NSL did NOT use aperture photometry at first was that the PSFs of the CONCAM2s was just so wide that software apertures could usually not be found that there free of the other stars and/or a sloping background.
The CONCAM3s have much smaller PSFs, though, likely making aperture photometry doable. Here is a good link from the American Association of Variable Star Observers (AAVSO) that has a good discussion of CCD aperture photometry: http://www.aavso.org/observing/programs ... al/4.shtml
But should NSL do aperture photometry? What value would it have for us? On the postive side it would likely encorprate more counts than even C25 for many stars, lowering statistical noise, and even pixel noise that bothers both C1 and C5. On the negative side, the extra outlying pixels likely have more error -- both statistical and systematic -- making it possibly less accurate than C25 or C16.
One possibility is that aperture photometry would better allow NSL data to track stars as they rise and set, which could mean the more precise use of comparison stars and the abilty to correct observed counts by a theoretical atmospheric absortion corrrection. I remain skeptical, though, that aperature photometry would really make a big difference here. It will never be able to compete with the accuracy of differential photometry between observations at the same sidereal time on different nights, in my estimation.
- RJN