Transcript Floating Windows for Webb Target Acquisition

Floating Windows for Webb
Target Acquisition
Mike
Regan


The simplest method for
locating a star is to determine
the 1st moment (centroid).
imax j max
Xc    i  Flux[i, j]
imin j min
imax j max
Yc    j  Flux[i, j]
imin j min
A star centered in the box has
an equal amount of light on all
sides.
As a star moves up in the box
there is more light below than
above -> 1st moment is wrong.
The first moment does not move as far as the star moved.
There are four variations on
using 1st moment for target
location.
• Raw 1st moment.
– One pass
• Gaussian weighted first moment
– iterative
• Bias-corrected first moment
– One pass
• Floating window first moment
– iterative
A simple method is to move the
box to always be centered on
the star. (Floating Window)
We weight the pixel fluxes by the ratio of the area of the
pixel in the box.
Bias in location is proportional
to observed location.
The bias-corrected and floating window have the
lowest RMS of the four methods (5 pixel box).
Raw 1st moment
Bias-corrected 1st moment
Floating window 1st moment
Gaussian weighted 1st moment
The Gaussian weighted method approaches the
bias-corrected and floating window for a 7 pixel
box.
Raw 1st moment
Conclusions
• The floating window will be the baseline
for all the instruments
– No calibration
– Simple logic
– High accuracy
• Gaussian weighted 1st moment has no
advantage over the floating window.