Wavelet-based image decomposition method for NuSTAR stray light background studies

Abstract

The large side aperture of the NuSTAR telescope for unfocused photons (so-called stray light) is a known source of rich astrophysical information. To support many studies based on the NuSTAR stray light data, we present a fully automatic method for determining detector area suitable for background analysis and free from any kind of focused X-ray flux. The method's main idea is `a trous' wavelet image decomposition, capable of detecting structures of any spatial scale and shape, which makes the method of general use. Applied to the NuSTAR data, the method provides a detector image region with the highest possible statistical quality, suitable for the NuSTAR stray light studies. We developed an open-source Python nuwavdet package, which implements the presented method. The package contains subroutines to generate detector image region for further stray light analysis and/or to produce a list of detector bad-flagged pixels for processing in the NuSTAR Data Analysis Software for conventional X-ray analysis.