Definition of roundness parameter

[asteroidguy]

How exactly is the roundness parameter defined that AIJ includes in its measurement table?

Thanks!

[karenacollins]

The reported values are a relative measurement of the roundness (vs. eccentric due to tracking or guiding problems) of the stellar PSF in the aperture. The measurement is a by-product of the centroiding routine which provides estimates of the FWHM values in both the x-axis and y-axis directions. The roundness is then 1 - ecc, where ecc = ((xWidth-yWidth)*(xWidth-yWidth)+4.0*mxy*mxy)/((xWidth+yWidth)*(xWidth+yWidth)), and where mxy is a representation of the combined FWHM estimate. If you need more detailed information, you'll need to look at the AIJ code in Centroid.java starting at line 540.

Karen

[karenacollins]

I'll also note that the FWHM estimate from the centroid routine is not particularly accurate. The "Seeing profile" (radial profile) algorithm provides a much better and direct FWHM measurement, but at the moment doesn't provide the individual X and Y axis direction FWHM measurements, so the "roundness" is not extracted from the radial profile.