FEASIBILITY OF HELIOSPHERIC IMAGING FROM NEAR EARTH

Abstract

Imaging solar wind structures via Thomson scattered sunlight has proved important to understanding the inner heliosphere. The principal challenge of heliospheric imaging is background subtraction: typical solar wind features are fainter than the zodiacal light and starfield by 2-3 orders of magnitude. Careful post-processing is required to separate the solar wind signal from the static background. Remnant background, and not photon noise, is the dominant noise source in current STEREO data. We demonstrate that 10× shorter exposure times would not strongly affect the noise level in these data. Further, we demonstrate that current processing techniques are sufficient to separate not only the existing background of the STEREO images but also diffuse variable backgrounds such as are expected to be seen from low Earth orbit. We report on a hare-and-hounds style study, demonstrating blind signal extraction from STEREO/HI-2 data that have been degraded by the addition of large-scale, time-dependent artifacts to simulate viewing through airglow or high-altitude aurora. We demonstrate removal of these effects via image processing, with little degradation compared to the original. Even with as few as three highly degraded source images over 48 hr, it is possible to detect and track large coronal mass ejections more than 40° from the Sun. This implies that neither the high altitude aurora discovered by Coriolis/SMEI, nor airglow effects seen from low Earth orbit, are impediments to a hypothetical next-generation heliospheric imager in low Earth orbit; and also that post-processing is as important to heliospheric image qualitiy as are optical contamination effects.