Imaging of the internal structure of an asteroid analogue from quasi-monostatic microwave measurement data

Abstract

Context. The internal structures of small solar system bodies (SSSBs) are still poorly understood. In this paper, we find an experimental tomographic reconstruction of coarse high-contrast details inside a complex-structured target object using multipoint full-wave radar data. Aims. Our aim is to advance the development of inversion techniques to be used in potential planetary scientific radar investigations targeting SSSBs, which have complex shapes and whose internal structure is largely unknown. Finding out the structure is an important scientific objective of Solar System research in order to understand its history and evolution. Methods. This is the second part (Paper II) of a joint study considering the methods to analyse and invert quasi-monostatic microwave measurement data of an asteroid analogue. We focused on incorporating advanced, full-wave, forward simulation in time domain with experimental data obtained from multiple measurement points. In particular, this study investigates multiple scattering and multipath effect suppression (MES) to reduce artefacts in the reconstructions. MES is necessary since the high-contrast and complex-shaped target and, especially, its back wall in high curvature regions cause intense reflections that deteriorate the reconstruction quality if not treated correctly. We considered the following two approaches to obtain MES: (i) geometrical optics-based pathlength thresholding and (ii) a peak detection method to investigate whether a data-driven approach could be used. At the inversion stage, we investigated marginalisation of random effects due to modelling by splitting a larger point set into several sparse sets of measurements. Results. Based on the results, MES is crucial to localise a void inside the complex analogue target. A reconstruction can be found when the maximum signal propagation time approximately matches that of the first back-wall echo for each measurement point. The marginalisation approach allows us to find a reconstruction that is comparable in quality to the case of full data, while reducing the computation effort per subsystem, which is advantageous when inverting a large data set.