High Resolution Imaging
I have led multiple research projects (20+ years) on high-resolution imaging of space objects, which is an integral component of astronomy and SDA. My research findings have offered cost-effective and simpler ways to reach the goal of producing improved image resolution. For SDA, high-resolution imaging enables the detection and characterization of space objects. Gathering information about the target’s shape and surface material composition is important for making inferences about its operational status and mission objectives. However, acquiring ground-based high-resolution imagery requires observing space objects through atmospheric turbulence using large-aperture telescopes, which adds additional costs and overheads that slow down research efforts.
I led the team that successfully delivered a daylight imaging capability to the USAF for SDA by deploying the Daylight Object Restoration Algorithm (DORA) at the Maui Space Surveillance Site (MSSS). Daylight imaging dramatically increases the volume of space and time to enable the observation and study of targets. However, atmospheric turbulence severely reduces the spatial resolution of imagery obtained by large telescopes. So, this requires wavefront sensing, adaptive optics systems, and advanced image restoration algorithms to obtain high-fidelity images. My research focuses on how to achieve this objective. Figure 1 shows examples of imagery and processed imagery obtained using the AEOS 3.6m telescope at MSSS during the evening terminator. Panel 1 shows an image of the Low-Earth-Orbiting Satellite as observed from the telescope (the turbulence blurs the image). Panel 2 shows restored imagery that I obtained using an image restoration algorithm that models the data using a turbulence model for the atmosphere. Another area of my research focuses on measuring turbulence effects on the imagery using a wavefront sensor (WFS). I used these measurements in the image restoration algorithm. This additional WFS data improves the data model, yielding further improvements in image quality (as shown in panel 3). Finally, the fourth panel depicts an image I obtained using the adaptive optics (AO) system, which utilizes the WFS measurement with a deformable mirror to compensate for turbulence-induced noise. Typically, this AO correction enhances the image’s resolution dramatically. However, my inclusion of the WFS data in the image restoration process has improved the image’s resolution, that is comparable to the AO-corrected image, but more cost-effective and simpler to execute.
Figure 1 Imagery from the Maui Space Surveillance Site (Hope, et al. 2018) From left to right: Raw Image (lower spatial resolution due to atmosphere), MFBD restoration (model of atmospheric PSF and target), MFBD restoration using WFS data, and AO-corrected image (Note that the satellite pose changes at the time during its pass when the AO is applied)
Another aspect of my research focused on imaging satellites in full daylight. Achieving high-resolution imaging during daylight is a daunting challenge due to the strength of the turbulence, which can be 10x the level experienced during terminator observations (see Figure 2 Panel 2). I overcame this challenge by building a multi-layer turbulence model, which increased the spatial sampling of the wavefront sensor. This allowed me to more effectively model the data and obtain high-quality images of satellites during full daylight. Please see an example of this daylight imaging of the ENVISAT remote sensing satellite obtained using the AEOS Telescope and the DORA algorithm is presented in panel 3 of Figure 2, where the left panel depicts the location of the telescopes at the MSSS site (AEOS 3.65m is the telescope with a silver dome).
Figure 2 From left: Maui Space Surveillance Site (AEOS telescope has silver dome), raw image of ENVISAT obtained during broad daylight, DORA image restoration with satellite components visible, and the artist rendition of ENVISAT and physical parameters. (Hope, et al 2018)
The deployment of the DORA algorithm was the culmination of many years of work in imaging through turbid media, image restoration, wavefront sensing, and adaptive optics systems. During the algorithm’s deployment at MSSS, I collaborated extensively with government sponsors and Boeing, the prime contractor on-site. This experience gave me a dual perspective on basic and applied research, highlighting how they complement one another. I have continued to develop these findings and am currently working with AFRL at MSSS to make enhancements to the DORA algorithm.
Advanced Image Restoration for Astronomy
I am the architect of the Kraken image restoration algorithm, which achieved the highest resolution of images of ‘Io’ ever obtained by a ground-based telescope. This unprecedented image quality rivals that of spacecraft images obtained by the Voyager 1 and 2 missions and has opened the opportunity to study geological events on Io from ground-based observatories.
This Kraken algorithm was applied to multiwavelength AO-corrected images of Io, captured using V, R, and I astronomical filters by the SHARKVIS Imager on the LBT in January 2024. The first three images are captured with the V, R, and I astronomical filters and show the image quality after AO correction. Please see the image on the right that shows the composite image of Io that I restored using the imagery from each of the three optical bands.
Figure 3 From left: V, R, I band images of Io captured using the SHARKVIS Imager (Italian National Institute for Astrophysics at the Rome Astronomical Observatory) on the LBT Telescope( Univ. of Arizona). The image on the right shows the composite image (compensated for Earth rotation) after image restoration.
Without this algorithm, future observing campaigns could not focus on studying surface activity and volcanic jets. Based on my team’s results, several planetary geologists plan to observe Ganymede and Europa. We received substantial media coverage for this achievement. As such, I have been invited to participate in a proposal to build new instrumentation for coherent imaging on the University of Arizona Large Binocular Telescope to advance the imaging of astronomical objects.
Searching for hidden Stellar Companions
I led a GTRI project (funded by NASA Ames Research Center) to improve detection capabilities for faint stellar companions using Speckle Image Data collected on the Gemini North and Gemini South 8m-telescopes. My work led to a new MFBD algorithm for speckle imaging and detecting faint stellar companions, demonstrating superior performance over conventional speckle interferometry. My algorithm produced significant improvements in binary/secondary flux ratio and higher precision astrometry, and they achieved contrast levels of 10-3 near the diffraction limit (20 milli-arcseconds) of the telescope and 10-4 at one arc-second separation. Please see Fig 4 panel-1 (speckle interferometry) and panel-2 (MFBD restoration) for Alpha Com and panel-3 (speckle interferometry) and Fig 5 panel-4 (MFBD restoration) for TOI-884.
Figure 4 From left: Alpha Coma Berenices (auto-correlation function), Alpha-Com (MFBD restoration), TOI-884 (auto-correlation), TOI-884 (MFBD restoration). Panels (1) and (3) show the mean autocorrelation of the image sequence. The sidelobes denote the signal of a companion. The MFBD image restorations in Panels (2) and (4) show the images of the binary system recovered using MFBD with the same image sequence.
This research has led to a new project – creating a new image restoration pipeline for Speckle Imaging. In the near future, I plan to lead the Speckle Imaging Program with NASA AMES, which will investigate archival data and new data with greatly improved flux ratio and astrometry estimates.
Image Deconvolution on JWST MIRI Imagery
I successfully applied the Kraken Image Restoration Algorithm to imagery obtained from the James Webb Space Telescope (JWST), by significantly increasing the science achieved on high-redshift Active Galactic Nuclei (AGNs) in the early universe. An AGN consists of a host with a central bright core that is magnitudinally brighter than the host galaxy itself. Due to the JWST primary mirror consisting of 18 hexagonal-shaped mirror segments, there is a significant number of sidelobe structures in the PSF, which is convolved with the image of the AGN. This reduces the image quality and limits the ability to characterize the core (see panel 1 in Fig. 5). My research focuses on removing this sidelobe structure from the image through deconvolution. Current deconvolution techniques perform poorly when modeling the bright core on the diffuse host galaxy. To overcome this limitation, I designed a new model for deconvolution of the AGN imagery that successfully reconstructs both the core and the smooth background free of sidelobe structure (see panel 2 in Fig. 5).
Figure 5 Left: JWST/MIRI composite image and Right: Deconvolved image showing AGN core and galaxy structure. (Leist, et al 2024) Images represent the composite image (by combing observations at 5.6, 10, 15, 18, and 21-micron wavelengths) of an AGN in galaxy NGC5728 obtained by JWST. The JWST PSF sidelobe structure has corrupted the AGN (center of the image) and blurred the host galaxy.
Importantly, I am leading an effort with our collaborators at the University of Texas and the NASA Space Telescope Science Institute to deploy this algorithm and new enhancements to create an imaging toolbox of algorithms for the JWST community.