Optimal Image Subtraction (OIS)¶
OIS is a Python package and a C command-line program to perform optimal image subtraction on astronomical images.
It offers different methods to subtract images:
- Modulated multi-Gaussian kernel (as described in [alard1998])
- Delta basis kernel (as described in [bramich2008])
- Adaptive Delta Basis kernel (as described in [miller2008])
Each method can (optionally) simultaneously fit and remove common background.
Theoretical Summary¶
All of the methods assume we have a reference image \(R\) and a science image \(I\) that can be approximately modelled as:
\[I \approx R \otimes K + B_{kg}\]
for some background \(B_{kg}\) and some kernel \(K\).
The optimal image subtraction \(D\) is then:
\[D = I - (R \otimes K + B_{kg})\]
The methods differ in their modelling of \(K\).
Warning
In the ideal case of perfect subtraction, \(D\) should contain only noise and optical transients. In practice, tiny image misalignments, saturated stars and poor PSF fitting can leave subtraction artifacts near sources.
[alard1998] | “A Method for Optimal Image Subtraction” - C. Alard, R. H. Lupton, 1997. |
[bramich2008] | “A New Algorithm For Difference Image Analysis” - D.M. Bramich, 2008. |
[miller2008] | “Optimal Image Subtraction Method: Summary Derivations, Applications, and Publicly Shared Application Using IDL” - J. PATRICK MILLER et al., 2008. |