Astro-Interferometric Modeling and Spatio-Spectral Reconstruction

NASA’s astrophysics roadmap deems interferometry necessary to achieve future science goals, particularly when the desired spatial resolution will require a telescope that is so large that it would be difficult to fabricate and test, too heavy to launch, and too costly. Interferometry can overcome these issues and provide high spatial resolution at the cost of lengthier data collection times. In addition to high spatial resolution, astrophysicists require spectral information to gain a better understanding of the material composition of the sources being imaged. Spatio-spectral interferometry combines the aperture synthesis aspect of Michelson stellar interferometry with the spectral measurement capabilities of Fourier transform spectroscopy, making it a great candidate for future space-borne interferometers. In order for spatio-spectral interferometric imaging to be a viable technique for future space missions, it must first be well understood. Our work on wide-field spatio-spectral interferometry is meant to advance and probe the technique, and bring it closer to technology readiness level 6. We have been developing algorithms that allow us to perform validation on simulated data as well as experimental data from NASA Goddard’s state-of-the-art Wide-field Imaging Interferometry Testbed (WIIT), which resides in the well-controlled environment of the Advanced Interferometry and Metrology laboratory. We fully developed the theory behind spatio-spectral interferometric measurements and the associated image synthesis algorithm. We used this knowledge to create a method of simulating the interferometric measurements and demonstrate the image synthesis algorithm on simulated data. Although image synthesis had been demonstrated prior to our work, we showed how the effective optical transfer function affects recovered spectra in addition to showing that deconvolution can be used to recover spectra that are more consistent with the original source spectra. We also contributed to the advancement and understanding of WIIT. We performed a phase retrieval experiment to recover the underlying aberrations imparted by WIIT’s imaging camera. We also developed an algorithm that allows WIIT’s scene projector, the calibrated hyperspectral image projector (CHIP), to display more complicated hyperspectral scenes than had previously been produced by CHIP for WIIT datasets. We also improved the preprocessing of WIIT data. We created an image registration algorithm that could recover unknown translations and rotations by combing the chirp z-transform algorithm with nonlinear optimization. The same algorithm could also perform deconvolution using our phase retrieval results, allowing for improved registration accuracy. The chirp z-transform was used to reorient all of the WIIT measurements to a single reference frame. We also developed a procedure for removing known systematic errors from the measurements to improve the overall quality of WIIT data. Finally, we developed a phase referencing algorithm that allows us to relate WIIT data taken at different baseline lengths and orientations, which is essential for performing image synthesis on experimental data. Throughout the fellowship, we had difficulty synthesizing testbed datasets. We discovered that there were unanticipated signatures in the data that we attributed to our difficulty. Over the last month we determined that the unanticipated signal features were time varying. From this realization, we found that the detrimental features of the signal are attributable to the source projector CHIP and fluctuations in the intensity of its light source throughout data collection. We provided possible solutions to this problem. Ideally, we should have a source that is more stable. If that is not possible, another solution would include monitoring the source intensity during data collection with a frequency at least as high as CHIP’s frame rate so that the unwanted time-varying signal could be processed out of the data.