November 18, 1999 Pratt Conference Room, Perkin G-04, 12:30 pm
Optical and Infrared Astronomy Division Talk
Microlensing Surveys and the Image Subtraction Method
Dr. Christofe Alard, Institut d'Astrophysique de Paris
The difficulties encountered in the analyzis of the microlensing
data motivated the developement of the image subtraction method.
Basically, the lensing optical depth and the lensing rates are very
affected by the contribution of a large number of faint unresolved
sources.This bias due to the faint stars operate by blending effects
which are very well corrected by the image subtraction method.
The basic of the image subtraction method is to match 2 images
by applying a convolution kernel, in order that the 2 images subtract
perfectly. Significant residuals in the subtracted image will appear
only at the position of the variable objects. The problem
of finding the best convoltuion kernel can be solved numerically
with minimum computing time. The method has been recently extended to
space-variant kernel, and has shown the ability to adress a large
number of problems in addition to the initial microlensing issue. These
other topics includes High-Z supernovae search, variable stars in
globular clusters, and also more original applications, like the
detection of T-tauri stars in nebulosities by subtraction R-band and
H-alpha images. Thus this optimal image subtraction technique is a new
and general method which can be useful and should be used in a number
of astrophysical applications. I will conclude by introducing the
ISIS image subtraction package.