深空攝影的 FWHM vs HFD
發表於 : 2012-01-30, 23:36現時建好西藏天文台採用自動化對焦系統，以HFD值為基準，跟傳統的FWHM值定義不同，剛看到一篇文字解釋這點：
----------------------------------------------------------------------------------------------
HFD Question
Posted by: "Jeff Woods" jeff@telix.com adbjester
Fri Jan 27, 2012 12:04 pm (PST)
Hello Gordon and Fernando,
OK, let's first set the wayback machine back before the days of
FocusMax. Back then, we measured stars in "FWHM", i.e. "full width,
half maximum". Full Width is the diameter of a star (in either pixels
or arcseconds). Half maximum means the distance from the center at
which the brightness drops from the peak value at the core, to one half
that max value.
Thus, FWHM means that if the centroid of a star is measured at 45000
ADU, the FWHM is the distance (in pixels or in arcseconds) that the
average ADU is 22500.
That measurement works well on stars of most any size or saturation and
most any pixel scale. It is more useful for comparing seeing conditions
when stated in arcseconds, and anything under 1.5 arcseconds FWHM is
very good seeing indeed.
Steve and Larry came up with the alternative, HFD. Unlike FWHM, HFD
will vary along with the overall flux (luminosity) of a star. Flux can
be thought of or measured as the sum total ADU of all pixels that make
up the star, minus the background sky noise. To simplify, if a star is
5 pixels by 5 pixels, and the total ADU count of all 25 pixels is
360000, you want to find the location of the circle that is centered on
the centroid, and that puts 180000 ADU inside the circle, and 180000
outside the circle. The diameter of that circle is the HFD.
While both measure *close* to the same thing, HFD works better for
focusing because it handles out-of-focus "donuts" far better -- the
measurements are linear except within the CFZ, whereas FWHM is not.
This makes V-curves possible, and thus makes auto-focusing via HFD more
robust than trying it with FWHM. HFD also takes seeing and noise into
account better than FWHM.
Jeff
----------------------------------------------------------------------------------------------
HFD Question
Posted by: "Jeff Woods" jeff@telix.com adbjester
Fri Jan 27, 2012 12:04 pm (PST)
Hello Gordon and Fernando,
OK, let's first set the wayback machine back before the days of
FocusMax. Back then, we measured stars in "FWHM", i.e. "full width,
half maximum". Full Width is the diameter of a star (in either pixels
or arcseconds). Half maximum means the distance from the center at
which the brightness drops from the peak value at the core, to one half
that max value.
Thus, FWHM means that if the centroid of a star is measured at 45000
ADU, the FWHM is the distance (in pixels or in arcseconds) that the
average ADU is 22500.
That measurement works well on stars of most any size or saturation and
most any pixel scale. It is more useful for comparing seeing conditions
when stated in arcseconds, and anything under 1.5 arcseconds FWHM is
very good seeing indeed.
Steve and Larry came up with the alternative, HFD. Unlike FWHM, HFD
will vary along with the overall flux (luminosity) of a star. Flux can
be thought of or measured as the sum total ADU of all pixels that make
up the star, minus the background sky noise. To simplify, if a star is
5 pixels by 5 pixels, and the total ADU count of all 25 pixels is
360000, you want to find the location of the circle that is centered on
the centroid, and that puts 180000 ADU inside the circle, and 180000
outside the circle. The diameter of that circle is the HFD.
While both measure *close* to the same thing, HFD works better for
focusing because it handles out-of-focus "donuts" far better -- the
measurements are linear except within the CFZ, whereas FWHM is not.
This makes V-curves possible, and thus makes auto-focusing via HFD more
robust than trying it with FWHM. HFD also takes seeing and noise into
account better than FWHM.
Jeff