Modern high-angular resolution ground- and space-based
imaging facilities have produced unprecedented views of
astronomical objects, from nearby stars to distant
AGNs and quasars. The images, obtained at wavelengths
ranging from gamma-rays to radio, contain many complex
emission components with different spatial scales and
with a wide range of contrast levels. It is therefore
difficult to detect, study, and model the multiscale
structures in these images and especially for low-count
sources.
The spatial resolution is often degraded because of
the blurring caused by the atmosphere and/or by the
telescope/instrument PSF. Modeling the PSF for each observed
source is the first step in the multi-scale analysis since
the size and the shape of the PSF may change significantly
as a function of time, off-axis angle, and wavelength.
Restoring the degraded resolution is even more challenging in
observations where the number of detected photons is
very low.
We describe here techniques and algorithms that allow
accessing multi-scale structures in low-count statistics
images. These include techniques for modeling telescope and
detector PSFs, adaptive smoothing, and deconvolution. We will
show examples of several tools and analysis threads, including
application to Chandra low-count statistics images and to other
observations of astronomical sources, and discuss current problems and future development.
Copyright. All Rights Reserved.