| |
 |
| Paper: |
The Characterization, Subtraction, and Addition of Astronomical Images |
| Volume: |
371, Statistical Challenges in Modern Astronomy IV |
| Page: |
160 |
| Authors: |
Lupton, R. |
| Abstract: |
Astronomical images are characterized by a spatially-varying Point Spread Function (“blur”; PSF) which is usually not known ab initio. Knowledge of the PSF is crucial for the correct analysis of images (e.g. separating stars from galaxies; estimating the deconvolved shapes of objects), and we have ways of modelling the PSF, but are not currently solving the problem especially well — the fields may be (very) crowded; the data may be undersampled; and the spatial variation of the PSF may be considerable (in the case of X-ray data, varying in width by an order of magnitude). A related issue is that of finding what has changed between a pair of images taken under different conditions. Rather than solve for the PSF of both images, we generally solve for the convolution kernel that transforms one into the other, but the question of how to represent this spatial structure of this kernel is similar to the (unsolved) problem of the previous paragraph. Finally, we are often faced with the question of how to handle a set of several to many images of the same part of the sky taken under different conditions, and which are, in general, marginally sampled. A variety of ad-hoc solutions exist to recover the properties of the objects in the field, but I do not believe that we currently solve the problem optimally. |
|
|
|
|
|
