Unconventional Imaging, Sensing, and Adaptive Optics 2024

Our work is the first to show via simulations based on experimentally measured, spectrally resolved AB stacked bilayer graphene pixel responsivity that complex spectral signatures can be reconstructed from 650 cm-1 to 750 cm-1 using fewer than 25 bilayer graphene detectors. Spectra reconstructions were performed using an Elastic Net algorithm that solves a least-squares minimization problem with additional L1 and L2 norm penalty terms. Then, we show for the first time that a single bilayer graphene pixel cooled to 80K can be used to detect atmospheric concentrations of CO2 either by direct evaluation of the pixel’s photocurrent or through the Elastic Net reconstruction technique, the latter of which relies on previous characterization of the detector’s responsivity as a function of gate voltage and wavelength.

We developed a polarimetric imaging drone for field inspections of CSP Heliostats. By utilizing the polarization pattern of the skylight calculated with the Rayleigh scattering model, the difficult scenarios for visible images show a good enhancement in the contrast of features, such as edges and cracks of the heliostat mirror facets. Analysis of the test results carried out at Sandia NSTTF validated the feasibility of applying this method and system to the CSP fields. Future work is desired for image fusion of high-resolution visible images and polarization images with well-designed angles.

Space Domain Awareness (SDA) is important for understanding the space environment to ensure safe operation of space missions. SDA activities include the detection, identification, tracking and characterizing of artificial satellites. SDA objectives rely greatly on the information that can be gained from ground-based sensors, such as optical telescopes. A space object may be detected by a telescope when it passes in front of a star. There have been studies of how the shadow cast to Earth from a star can be interpreted for important data, referred to as Shadow Imaging. Herein we discuss the usage of information theoretic methods to understand the limitations of stellar Shadow Imaging. These methods measure the information content in the irradiance pattern, as seen by a terrestrial observer, from the shadow cast by the geosynchronous space object passing in front of a stellar source.

Estimating the silhouette of large space objects through decreased intensity measurements is an established technique in astronomy. This technique has been expanded to satellites near Earth in mathematical and computer models as well as scaled laboratory demonstrations. Synthetic Aperture Silhouette Imaging (SASI) applies these concepts using a North-South oriented linear array of hobby telescopes to detect decreased intensity from stars as satellites occult the stars. This paper discusses the challenges and successes of an initial field test using a single telescope equipped with a photon detector. The goal is to measure intensity drops from stars when satellites are predicted to pass between the star and the telescope in a ground station. This initial field test will inform the design of an individual telescope in a SASI array and validate the use of the selected hardware as an initial step towards the design of a full SASI array.

This work presents an innovative approach to simulating image formation in the human eye by developing a bionic eye model. Commercial software is employed to create an accurate model of the human eye for simulating eye images. Accordingly, A physical bionic human eye model is crafted to explore the imaging properties of eye images and validate the image quality of the eye model, with a liquid lens embedded as the crystalline lens for accommodating the eye at different object distances. To comprehensively assess the performance of the bionic eye model, digital holography is utilized to measure and analyze the wavefronts of the eye model to understand the aberrations of the bionic human eye and their impact on image quality. This study provides a thorough evaluation of the performance of the proposed eye model to enhance the accuracy of bionic eye imaging.

Coherent diffraction imaging is a lens-less microscopy technique that only measures the intensity distribution of the diffractive pattern. The phase is normally retrieved by iterative algorithms. In this paper, we propose a new approach of phase retrieval algorithm in coherent diffraction imaging for flat optics inspection. We use the phase of the designed flat optics as reference and combine with the measured diffractive intensity of the fabricated flat optics. In this way, we obtain higher resolution images at a faster rate comparing to the conventional blindly iterative algorithm. This proposed method will benefit flat optics inspection in the aspects of high resolution, compact imaging set-up, and efficient phase retrieval algorithm.

We design large aperture, all-silicon meta-optic doublets for unidirectional and synergistic imaging at a wavelength of 4 μm. When illuminated by a plane wave in the forward mode, our unidirectional imager generates an intense spot on its optic axis at a predefined focal length. In the reverse mode, the imaging performance is significantly reduced, accompanied by a dramatic reduction in light intensity on the focal plane. On the other hand, our synergistic imager is optimized to focus an incoming plane wave only when its constituent meta-optics are used in conjunction with each other. We envision our devices to provide new avenues for the development of metamaterial imaging platforms for applications in defense and data security.

This talk examines the potential of polarized microscopy techniques utilizing Mueller Matrix formalism and image decomposition in detecting skin melanoma. Unlike the "gold standard" dermoscopy technique which has low accuracy in detecting melanoma and is unable to reveal collagen structure in the skin extracellular matrix, our system is simple, portable, allowing real-time screening in the physician’s office, and can highlight collagen features.

ROCS microscopy, a label-free super-resolution technique, provides high spatial (150 nm) and temporal (10 ms) resolution, ideal for live-cell imaging. By rotating a blue laser beam at large angles, ROCS captures interference patterns, yielding high-contrast images without post-processing and minimal laser power. This method allows rapid acquisition of thousands of images of live cell dynamics. In the ROCS Bright-field (BF) mode, coherent amplification of scattered light enables imaging at lower optical powers, facilitating prolonged tracking of cellular interactions. Utilized for Fibroblast-born tunneling nanotubes (TNTs), which play a vital role in cardiac cell communication, ROCS employs total internal reflection (TIR) and dark-field (DF) modes to measure TNT growth with exceptional contrast. With the capability to reach depths of up to 5 micrometers in Non -TIR mode, ROCS effortlessly switches between modes, establishing itself as a versatile tool to study biological samples.

This work demonstrates the impact of sampling basis resolution on Single Pixel Imaging's ability to accurately measure the amplitude and phase of an arbitrary complex light field. We observe that phase crosstalk occurs in the reconstructed amplitude when the sampling basis is coarser than the underlying phase distribution. To investigate this phenomenon, a spatial light modulator is employed to prepare the complex field—comprising both amplitude and phase—utilizing a phase-only hologram. The field is then sampled at various resolutions using the Hadamard basis on the same SLM with interferometric Single Pixel Imaging. Measurement of the field takes place on a single point through 3-step interferometry and a Mach-Zehnder interferometer. While the phase is correctly reconstructed for each resolution, our findings reveal distinctive attributes of the phase distribution in the reconstructed amplitude for coarser sampling resolutions.

Active imaging techniques such as continuous-wave (CW) illumination and laser range-gated (LRG) imagery can provide increased contrast-to-noise ratio over passively-illuminated systems. The development of an active imaging testbed with CW and LRG imaging capabilities in the shortwave infrared (SWIR) and extended SWIR (eSWIR) bands is discussed. The results of a field imaging study are shown and compared to radiometric performance modeling, and the agreement between these two approaches is discussed.

This work addresses the challenge of speckle noise in coherent imaging techniques for dynamic scenes, such as synthetic aperture radar, sonar, and holography. We propose an unsupervised deep learning method that effectively denoises time-varying sequences by fitting a densely connected neural network to sequential measurements. Our approach creates a time-varying implicit neural representation, optimizing the model by maximizing the log likelihood of reconstruction given a forward model of the measurement and noise process. This results in accurate estimates of the underlying scene reflectivity, even in the presence of spatially correlated speckle from a physical aperture. Comparative evaluations with state-of-the-art denoising methods demonstrate superior performance.

In coherent imaging systems like SAR and digital holography, speckle noise is effectively mitigated using the multilook or multishot approach. Utilizing maximum likelihood estimation (MLE), we recently theoretically and algorithmically showed the feasibility of effectively recovering a signal from multilook measurements, even when each look is severely under-determined . Our method leverages the "Deep Image Prior (DIP) hypothesis," which posits that images can be effectively represented within untrained neural networks with fewer parameters than the total pixel count, using i.i.d. noises as inputs. We also developed a computationally efficient algorithm inspired by projected gradient descent to solve the MLE optimization, incorporating a model bagged-DIP concept for the projection step. This paper explores the method's applicability to deblurring in coherent imaging, where the forward model involves a blurring kernel amidst speckle noise—a significant challenge with broad applications. We introduce a novel iterative algorithm to address these issues, enabling multi-look deblurring without simplifying or approximating the MLE cost function.

Pending public release approval.

Abstract awaiting AFRL opsec release.

Abstract pending OPSEC review

Abstract pending OPSEC review

The aero-optical distortions caused by supersonic mixing layers over a flat optical window are relevant to the performance of hypersonic vehicles. Such mixing layers are typically temperature-mismatched and gas species-mismatched due to a need to efficiently cool the optical window. To investigate the effect of the mismatched properties across the mixing layer created by blowing a cool gas over a flat window, optical measurements of an M=2 freestream air flow with a cooling two-dimensional gas jet were taken using time-resolved Shack-Hartmann WFS and Schlieren photography. To complement the optical measurements, other non-intrusive techniques, specifically acetone-based PLIF and spatially-resolved infrared thermography techniques were implemented to extract relevant fluidic properties of the mixing layer. Results of this experimental work will be presented and discussed. Parameters of the mixing flow were examined using optical velocity methods. Based on experimental input, various scaling methods proposed in previous work were implemented in order to predict aero-optical distortions.

Abstract pending OPSEC Review

Pending OPSEC approval

Light-sheet microscopy (LSM) is known for increased image contrast and reduced photo-bleaching since only the parts of the object are illuminated with laser beams from the side, which are in the focus of the objective lens. In addition, larger volumes are scanned plane-wise or line-wise, such that LSM is significantly faster than point-wise scanning methods. However, compromises in spatial resolution have had to be made because of objective lenses with limited numerical aperture and aberrations from light scattering. However, spatial light modulation and phase adaptation of the illumination side still leave plenty of room for improvement. In the first step, we tackle the problem of limited resolution by combining holographically shaped illumination beams (e.g. Bessel beams) with three-axis interferometric arrangements of the illumination beams. By generating interference fringes between every two beams from different illumination objective lenses, we use structured illumination microscopy, to obtain 3D super-resolved images in scattering media.

In conventional Light-sheet microscopy where fluorescence carries information about the optical properties of a sample, excitation light that is scattered through sample is often neglected. We investigate propagation of Gaussian and Bessel beams through classes of spheres, proposing a theory to extract scattering coefficients from analysis of the forward scattered light. These coefficients aid in detecting sample changes not perceivable from light-sheet images alone, providing position-dependent refractive index and density data in conjunction with standard fluorescence imaging. We examine the degradation of image quality along optical axes and the position-dependent scattering and absorption of illumination and fluorescence light on silica beads and cancer cell clusters.

Volumetric scattering in a disordered photonic media leads to a universal interference pattern called speckles, which poses significant challenge when probing passing through such a medium. To address this, here we report an interferometric based technique to modulate and probe optical fields inside a random photonic media. The method relies on quantification of monochromatic transmission matrix, whose statistical properties follows predictions of the random matrix theory. The knowledge of it also enables sensing of thermos-optical modulation of an optical field and essentially quantifying the phase distortion and temperature profile. The results have direct implication on different fields from imaging and information technology to biomedical optics.

Nonmelanoma skin cancers represent the largest proportion of all cancers, and their incidence is increasing worldwide. Past studies suggest that cancerous tissue have changed amounts of various naturally occuring cutaneous flourescent molecules when compared to normal tissue which can be studied using methods that detect autofluorescence. Autofluorescence photobleaching imaging and fluorescence lifetime imaging have been studied as a potential approach to early diagnosis of nonmelanoma skin cancer. To increase understanding of autofluorescence and autofluorescence photobleaching processes in different types of tissue these methods will be used to assess intensity and photobleaching rates in malignant and benign skin diseases. For autofluorescence photobleaching imaging a custom device was used that induces autofluorescence using 405 nm LED illumination and CMOS sensor for caputuring images. For fluorescence lifetime imaging the measurements were taken using a 405 nm laser for autofluorescence induction and a SPC-150 TCSPC module with a photon counting detector. Results will be used to further develop these methods for nonmelanoma skin cancer assessment.

Transient events exhibit strong UV radiation, but transient activity is not well studied in the UV. Ground-based telescopes have an untapped potential to support space-based UV observations of transients. The Super-LOTIS (Livermore Optical Transient Imaging System) telescope is the first ground-based optical telescope that is being converted for NUV transient science. It will follow up on transient targets identified by Swift’s UVOT instrument, future space-based missions, and conduct its own observations. Super-LOTIS will also provide a model for the future of ground-based UV transient surveys.

"abstract pending OPSEC release"

In previous work, it was shown that large density gradients adversely affect the wavefront reconstruction of Shack-Hartmann wavefront sensors (SHWFS). We also demonstrated that removing distorted dots from the Shack-Hartmann image achieves some improvement in the wavefront reconstruction. In this work, we continue investigating this problem, whereupon removing these distorted dots, we applied various interpolation methods to fill in the missing points. The resultant wavefronts are compared to wavefronts independently measured using the digital holography wavefront sensor (DHWFS). Various experiments were performed to generate shocks of different strengths in the transonic to supersonic flow regime, and the shock-affected wavefronts were simultaneously measured using the DHWFS and SHWFS. The distortions in these dots are related to the properties of the supersonic shock, such as shock strength, spatial extent, and location. Different approaches were investigated to study the specific conditions under which these distortions can occur. These results can be used to improve the Shack-Hartmann wavefront reconstruction algorithm in the presence of sharp gradients.

Abstract awaiting public affairs approval

Pending OPSEC review

Abstract pending public release approval.

Pending OPSEC review

In this work, we present simulated results comparing the performance of closed-loop adaptive optics (AO) driven by a Shack-Hartmann wavefront sensor (SHWFS) with that of an AO system using a four-sided pyramid wavefront sensor (PyWFS) under identical deep turbulence conditions. For direct comparison of the two wavefront sensors’ capabilities, the resolution of both were set to be 32 elements across the full pupil diameter: 32 lenslets across the SHWFS and 32 pixels across a single pupil of the PyWFS. Phase screens of varying Rytov numbers were generated by propagating a wavefront through multiple atmospheric layers defined by Kolmogorov statistics, resulting in both phase aberrations as well as scintillation. These phase screens were then used as the aberration source for both wavefront sensors to enable comparison of their performance under the same turbulent conditions. Simulated wavefront compensation was provided by a single-plane deformable mirror model conjugate to the wavefront sensor. The results of this work is a comparison of the wavefront correction achieved by SHWFS-driven AO and PyWFS-driven AO in terms of power-in-the-bucket (PIB) versus Rytov number.

A wave-optics-based numerical simulation analysis of the impact of speckle-beacon size on the performance of an adaptive optics (AO) system operating in volume atmospheric turbulence is presented. For clarity, the speckle beacon was represented by a laser beam of a super-Gaussian profile, scattered off an extended flat target with Lambertian surface roughness. The control loop of the AO system included a high-resolution scintillation-resistant multi-aperture phase contrast wavefront sensor (MAPCO WFS), an ideal (infinite-resolution) wavefront corrector, and phase-conjugation (PC) type controllers utilizing either conventional PC or advanced speckle-average (SA) PC control algorithms. The results obtained show that the use of the advanced control algorithms makes it possible to partially mitigate the target-induced speckle effects and turbulence-induced wavefront aberrations for extended beacons, the size of which is comparable to or even exceeds the diffraction-limited beam spot size of the corresponding laser beam projection system.

The objective of the research is to investigate the use of multi-conjugate adaptive optics for compensation of strong atmospheric turbulence. With multiple wavefront sensors and deformable mirrors placed specific conjugate planes along the beam propagation path, multi-conjugate adaptive optics aims to lessen the scintillation effects and the phase aberrations each wavefront sensor and deformable mirror needs to handle. To study the effects of multi-conjugate adaptive optics compensation compared to single-conjugate adaptive optics, a laboratory testbed with two deformable mirrors and two wavefront sensors has been developed at the Naval Postgraduate School. In this work, a detailed description of the laboratory multi-conjugate adaptive optics testbed is presented. Current experimental results on multi-conjugate adaptive optics compensation of laboratory simulated atmospheric turbulence are also presented.

The MMT Adaptive optics exoPlanet characterization System (MAPS) is currently in its engineering phase, operating on sky at the MMT Telescope. The MAPS Adaptive Secondary Mirror’s actuators are controlled by a closed loop modified PID control law and an open loop feed forward law, which in combination allows for faster actuator response time. An essential element of achieving the secondary’s performance goals involves the process of PID variable tuning. In this paper, we will discuss the actuator control system, our methodology for tuning the PID loops and creating the feedforward matrix, including the use of actuator positional power spectra for both tuning and determining the health of individual actuators. We will present the comparative results of running the ASM without tuning control, with tuning control, and with the feedforward system in place. We will conclude by presenting an overview of our next steps.

We present an improved method for high-speed, target-in-the-loop (TIL), adaptive optics (AO) with phased-array lasers that uses separate feedback loops for local and atmospheric phase correction to maintain coherent combination with fewer control degrees of freedom (DOFs) than previously required. TIL AO systems monitor and maximize back-reflected laser power from the target using a hill-climbing algorithm to iteratively adjust the phase within each of the N subapertures. Previously, because local phase can vary by many waves between subapertures, the number of DOFs needed to be at least N-1. This undesirably links the AO correction bandwidth with the number of subapertures, limiting power scaling for phased-array lasers. When local phase is corrected separately, spatiotemporal correlations in turbulence can be exploited to reduce the number of DOFs far below N-1 (e.g., by hill-climbing with an undercomplete set of Zernike modes). We use machine learning algorithms to construct optimal mode sequences for hill-climbing in different atmospheric conditions and use reinforcement learning to develop control schemes that select the best hill-climbing strategy based on real-time feedback.

Using digital holographic (DH) sensors, coupled with iterative computational algorithms we can sense and correct distributed volume turbulence in DH imagery. These iterative methods minimize a non-convex cost function with respect to the wavefront phase function, modeled as discreet arrays. This leads to high-dimensional optimization problems plagued by local minima. This problem is amplified in the presence of challenging conditions, (e.g., high noise, strong turbulence, insufficient data). We investigate using implicit neural representations (INRs) to model atmospheric phase errors in DH data. INRs offer a low-dimensional functional representation, simplifying the optimization problem and allowing us to produce a high-quality wavefront estimate.

In this paper we explore the use of digital holography to sense and correct atmospheric turbulence. Previous laboratory experiments at the United States Air Force Academy have demonstrated both sensing and correction of simulated atmospheric turbulence using an internally developed digital holography system. To expand the current state of the art in digital holography, we design and test a digital holography system in relevant atmospheric conditions in a field experimental campaign. Our results detail the performance of the digital holography system in an operationally relevant field test environment. PA#: USAFA-DF-2024-54

Imaging-based particle localization and tracking measurements can be subject to significant measurement errors if the measurement is performed through a temporally varying air-water interface. This situation occurs in a huge variety of technical energy conversion processes like bubble formation in electrolysis, droplet formation in fuel cells, or film flows. An actuator-free approach for the correction of time-varying low-order aberrations is presented. It is based a multiple-input deep convolutional neural network that uses an additional wavefront sensor input. Application is demonstrated by means of a flow measurement through an open, oscillating water surface. We show that the measurement error of the flow velocity induced by the fluctuating aberrations can be reduced up to 82 % if the correction is applied. This actuator-free approach has a potential to correct distortions in real-time which are uncorrectable for traditional AO systems which are limited by the performance of available actuators.

Wave optics simulation of atmospheric beam propagation with adaptive optics control loop is numerically simulated. Various combinations of wavefront sensors, including Shack-Hartmann and Digital Holography based wavefront sensors, along with adaptive predictive control is analyzed to determine optimal system level design for high speed adaptive optics for long distance beam propagation. Tradeoffs between improved accuracy of wavefront correction and additional latencies introduced from alternative system design choices is discussed. Pending public release approval

For several decades, the Tyler frequency has provided the tracking community with a reliable estimate of bandwidths required to consistently track an object through turbulence. That analysis, however, arrives at tractable solutions by operating in continuous rather than discrete time. Furthermore, it assumes only a first-order lowpass filter as the dynamic controller model. This paper extends Tyler’s original treatment to address each of these potential limitations, then proceeds to identify an additional bandwidth constraint from image-plane speckle noise associated with coherent illumination. The trade space poses an optimization problem, with proposed solutions in the form of modified bandwidth requirements that depend upon system diffraction angle, object angular velocity and speckle contrast ratio. Validation from wave-optics simulations and computer-aided control system design informs the analytical tools developed here and demonstrates their applicability to modern challenges in active tracking.

This paper explores the potential of using wide band super continuum lasers to improve link availability in turbulent atmospheric conditions. The analysis employs wave optics simulation and perturbation theory to assess the performance of the optical link. The focus is on reducing fade statistics and ensuring reliable and robust communication. The study considers various factors such as range, aperture diameter, wavelength, and turbulence levels. The paper also discusses the benefits of incoherent beam combining from spatial diverse laser sources in improving link availability, throughput, and stability in turbulent atmospheric conditions.

Atmospheric-induced amplitude fluctuations, known as scintillation, impose limitations on active tracking and wavefront sensing performance over near-horizontal propagation paths. These systems typically use centroid tracking to estimate the aperture-averaged phase gradient (G-tilt). In the presence of scintillation, the centroid tracking will measure the irradiance-weighted average phase gradient (C-tilt). Of particular interest are the effects of scintillation on the estimation of the centroid angular position. Scintillation will cause an error in the estimation of the gradient tilt and can be quantified by the noise-equivalent angle (NEA). This work formulates closed-form expressions for the NEA due to scintillation and mean-squared error between G-tilt and C-tilt in the weak-to-moderate scintillation regime. SMDC PAO#4010 6Feb2024

Pending public release approval

This talk addresses the challenges of tracking extended targets using optical tracking and the limitations of collecting real-world data for algorithm development. It proposes the use of synthetic imagery as a solution, which allows for customizable and easily accessible scenarios. Synthetic imagery can be a valuable tool for exploring tracking algorithms and designing robust systems but has limitations compared to real-world scenarios. This talk aims to define the capabilities and limitations of synthetic imagery and its impact on system evaluation.

The angular spectrum method is widely used in wave optics simulations, including simulations of light propagating through distributed-volume turbulence. A well-known variant of the method, the scalable angular spectrum, allows one to vary the sample spacing in the output plane for greater flexibility. This paper examines the scalable angular spectrum method in greater detail and describes how this method can be expressed as both two back-to-back Fresnel transforms or as a single chirp-Z transform. Additionally, we explore two possible interpretations of the outputs from the scalable angular spectrum method and relate the two interpretations via a nonlinear transform. Finally, we examine some shortcomings of the angular spectrum method when simulating distributed-volume turbulence and compare to the newly developed sinc-propagation method. The results indicate that phase screens generated throughout the propagation path can have adverse effects on output fields when using the angular spectrum method, and the sinc propagation-method is better suited for simulating these cases.

Wave optics simulations are a common method of studying the effect of beam propagation through the atmosphere. Current widely-used software utilizes phase screens that are statistically representative of Kolmogorov turbulence where the scattering parameter along the laser path relies on an atmospheric structure parameter which is typically measured using a path averaged scintillometer. While this method works for homogenous atmospheric conditions in which there is little variation between the transmitter and the receiver, in adverse environments such as those with high precipitation rates, heavy fog, or clouds, this method of modeling beam propagation can be insufficient. This work computes scattering from a “first principles” approach, where the full Mie series is calculated at several locations along the simulated beam path. The scattering results are then combined with traditional split-step beam propagation methods resulting in a more-complete, and therefore more-accurate, beam performance model. The simulation results are compared to experimentally measured statistical beam profiles, showing that including scattering models results in a more accurate prediction of performance.

We present a novel approach to split-step wave-optical simulations that are routinely used to simulate optical propagation through random inhomogeneous media. In this framework, the propagation of waves through the medium is computed through an alternating sequence of Fresnel diffraction integral computations and stochastic phase distortions. Current methods are limited to static scenarios and primarily rely on Fourier series approximations. Therefore, these methods are limited in their utility and suffer from numerical artifacts due to artificial periodicity. We develop a new approach to both computing the diffraction integrals and dynamically generating phase screens using sinc bases that are naturally suited to unbounded domains. The efficient computational methods that follow from this approach allow for arbitrarily long time-dependent simulations that compute the correct evolution of the physical two-point statistics of the medium.  We demonstrate the power of these methods in complex scenarios of practical interest to scientific applications.

Pending public release approval I submitted the abstract for AFRL approval in mid January. After getting AFRL approval, I need to get Lockheed Martin approval. I informed Matt Kalensky of this and will keep him informed.

Frodo is a new Differential Image Motion Monitor (DIMM) developed as a companion instrument to the Starfire Atmospheric Monitor (SAM). SAM is a Shack-Hartmann sensor that observes bright stars to provide atmospheric information above the Air Force Research Laboratory's Starre Optical Range (SOR). Frodo is designed to extend atmospheric characterization at the SOR into the day, estimating all of the atmospheric parameters that SAM currently provides. The optical train of Frodo has been reproduced on the Multiconjugate Adaptive Optics (MCAO) bench in the Atmospheric Simulation and Adaptive Optics Laboratory Testbed (ASALT) at SOR. The enhanced atmospheric turbulence simulator (ATS) on the MCAO bench generates turbulence conditions with as many as 10 phase screens. The atmospheric parameter estimates from data collected with the ASALT Frodo system is presented alongside the estimates made with the laboratory's Shack-Hartmann wavefront sensor.

Waiting for public release approval

The Energy Balance Bowen Ratio (EBBR) – the ratio of sensible heat flux to latent heat flux – has shown promising capabilities to calculate sensible heat flux, a key component for computing CT2 and Cn2. Sensible heat as calculated via the Bowen Ratio inherently accounts for moisture content and evaporation as it apportions the balance of sensible heat to latent heat in the ratio. Thus it better permits the calculation of CT2 and Cn2 via a single equation only dependent on temperature and sensible heat in any stability condition as compared to “ground truth” turbulence values during daylight and nighttime hours at various land sites. The Bulk Aerodynamic method relies on standard meteorological observations but requires stability corrections based on underlying assumptions. In this study, field data from a marine wave boundary layer test site allow for assessments of both the EBBR and Aerodynamic methods to quantify maritime surface layer turbulence.

The Small Mobile Atmospheric Sensing Hartmann (SMASH) provides a range of turbulence parameters for each data set collected. Typically the SMASH units collect 300 frames of data at 300 Hz once per minute to provide characterization of the atmospheric conditions over a path on which an adaptive optical system is being tested. Several algorithms used by SMASH estimate a turbulence parameter, the Fried parameter or the scintillation index for example, on a single frame and then average that value over all of the frames collected. In this work the estimates of the individual frame values are compared to the known turbulence conditions simulated with the reflective atmospheric turbulence simulator (RATS) in the Air Force Research Laboratory's Beam Control Laboratory.

pending public release approval

The reflective atmospheric turbulence simulator (RATS) in the Air Force Research Laboratory's Beam Control Laboratory is used to impart realistic distortions to a propagating wavefront for testing in the advancement of adaptive optical technologies. Here RATS is being used to simulate turbulence conditions over which the Small Mobile Atmospheric Sensing Hartmann (SMASH) system has typically operated. An optical clone of the SMASH system installed on the optical bench behind RATS measures the imparted optical disturbance and makes an estimate of the turbulence profile. The results are compared with the profile calculated based on the configuration of the RATS system.

In optical propagation experiments, knowledge of the turbulence along the optical path is key to quantifying the performance of the optical system. The Small Mobile Atmospheric Sensing Hartman (SMASH) units were designed to provide an estimate of a range of integrated atmospheric parameters used in validating the performance of adaptive optical systems. The latest upgrade to the SMASH capability is its ability to provide a profile of the index of refraction structure constant, C2n(z), along the test path. The differential scintillation technique has been studied in simulation and compared in field experiments against sonic anemometers over a horizontal path close to the ground. The later involved turbulence conditions that were beyond those for which the system was designed. Here testing is done along one of SMASH's normal testing paths at the Environmental Laser Test Facility (ELTF). The independent sampling of the atmosphere in this work is done using a drone equipped with an Optical Differential Temperature Sensor (ODTS) and flown along the optical path. The measurements of the two instruments are compared over the range of data collected.

Abstract awaiting public release review from AFRL/RD. Will update abstract with approval is received.

Recent studies have demonstrated that incorporating phase information alongside registered intensity distribution enhances Cn2 predictions by Deep Neural Networks (DNNs). In this study, we conducted numerical simulations to analyze the impact of received phase information on Cn2 recovery error, specifically examining the size and number of lenses required to form an intensity pattern in the focal plane. We compared the DNNs predictions performed by data obtained using different phase sensors, which were comprising various lenslet arrays. Furthermore, we investigated DNNs trained on images captured under varying Cn2 conditions within specific turbulence inner and outer scales, contrasting them with DNNs trained across a range of possible turbulence scale values.

Branch points are seen in many adaptive optical experiments where the sensed beam has propagated over an extended path, through sufficiently strong turbulence. It has been shown that branch points provide information on how the turbulence responsible for their formation is distributed and moving along the path. Shack-Hartmann wave front sensors have previously been somewhat limited in their ability to fully capture the branch points present within their measurements. A new technique for the detection of branch points based on the second moment statistics of the individual spots in images created with a Shack-Hartmann wave front sensor is examined. Data collected by Small Mobile Atmospheric Sensing Hartmann (SMASH) units are used to test the method under a range of turbulence conditions. The results of the second moment technique are compared with the standard elementary circulation method.

In this work, simultaneous SHWFS and DHWFS will be collected for a beam that propagated through benchtop-simulated atmospheric turbulence. Here, spatial light modulators (SLMs) will be used to represent turbulent phase screens and a laser beam will be propagated through the system designed to replicate distributed-volume atmospheric turbulence. The simultaneous SHWFS and DHWFS measurements will be collected for varying optical-turbulence conditions and the results will be compared. This research will inform efforts looking to perform adaptive optics in distributed-volume turbulence.

The reflective atmospheric turbulence simulator (RATS) generates turbulence for benchtop experiments for the Air Force Research Laboratory’s Beam Control Lab (BCL). RATS has six moveable reflective phase wheels, six moveable relay modules, and seven rails which can be used to configure the system to provide for a broad range of conditions under which to conduct tests. It is important for the BCL to have a well calibrated method of generating turbulence for any future aero-optical or beam control experiments. This paper explores the capabilities of RATS using Shack-Hartmann wavefront sensor (WFS) measurements. The turbulence parameters estimated from Shack-Hartmann data are presented alongside those calculated theoretically for different RATS configurations.

When an intense laser interacts with a gas of atoms, high-order harmonics are generated. In the time domain, this radiation forms a train of extremely short light pulses, of the order of 100 attoseconds. Attosecond pulses allow the study of the dynamics of electrons in atoms and molecules, using pump-probe techniques. This presentation will highlight some of the key steps of the field of attosecond science.

Measuring the fluid flow of facility effluents using event-based sensors (EBS) moves us from identifying whether a process is occurring to quantifying facility emissions, since no remote method for persistent monitoring of fluid flow currently exists. Effluents have proven to be a key signal in facility monitoring. Persistent monitoring of facilities requires sensors with low power and data bandwidth requirements, but characterizing the fluid dynamics of plumes requires high time response and dynamic range. EBS provide a compelling sensor solution due to their low power consumption (<1.5W), low data volumes, high dynamic range (10^6), and fast response times (20us). However, asynchronous/frameless EBS data do not allow for conventional image analysis and require new algorithms for motion characterization. Previous work demonstrated the possibility of using EBS to track complex motion of rigid bodies (e.g., people, drones, beads in fluids) with high temporal resolution, but measuring fluid flows remotely requires measurements of diffuse and dynamic edges (e.g., aerosols entrained in eddies). This has been demonstrated in imaging cameras, but not EBS.

Event-based vision sensor (EVS) technology has expanded the CMOS image sensor design space in terms of high-dynamic range operation and ability, under certain conditions, to efficiently capture scene information at a temporal resolution well beyond that achievable by a typical sensor operating near a 1 kHz frame rate. Fundamental differences between EVS and framing sensors necessitates development of new characterization techniques and sensor models to evaluate hardware performance and examine the trade-space between the two camera architectures. Here we present progress on development of basic EVS characterization capabilities and techniques for commercial-off-the-shelf visible event-based cameras, to include pre-launch characterization results for the two Falcon ODIN (Optical Defense and Intelligence through Neuromorphics) EBCs scheduled for launch to the International Space Station (ISS). Falcon ODIN is a follow-on experiment to Falcon Neuro previously installed and operated onboard the ISS.

Asynchronous, neuromorphic, event-based imagers have many advantages over traditional imagers (e.g., low power, efficient data collection, high speed, high dynamic range), on account of primarily detecting changes in log light intensity on a per-pixel basis. However, the temporal change detection nature of these imagers also introduces some challenges when trying to employ common camera calibration techniques such as Zhang’s method. In this work we propose the use of digital coded exposure techniques combined with blinking light emitting diodes (LED) to facilitate the process of camera calibration with asynchronous event imagers.

Falcon ODIN is a technology demonstration science payload being designed for delivery to the International Space Station (ISS) in late 2024. Falcon ODIN contains two event based cameras (EBC) and two traditional framing cameras along with mirrors mounted on azimuth elevation rotation stages which allow the field of regard of the EBCs to move. We discuss the mission design and objectives for Falcon ODIN along with ground-based testing of all four cameras.

Event-based Shack-Hartmann wavefront sensors are a promising sensing technology that can measure wavefronts at unprecedented framerates. However, they still suffer from some of the same limitations as conventional Shack-Hartmann sensors—severe turbulence can cause scintillation and strong wavefront distortions that corrupt the wavefront measurements. We propose a robust and computationally efficient deep learning framework for processing and reconstructing event-based wavefront measurements. We extend our approach to predictive wavefront sensing to further reduce system latency. We find our method can perform reliable wavefront reconstruction in the presence of severe time-varying turbulence.

Event-Based Shack-Hartmann Wavefront Sensor for Adaptive Optics Development and Testing

Nuclear Scene Data Fusion (SDF) integrates radiation data and contextual sensor inputs to enable free-moving localization and mapping of nuclear radiation from mobile sensor systems referred to as Localization and Mapping Platforms (LAMPs). The LAMP sensor suite typically includes lidar, an Inertial Navigation System (INS), and an RGB(D) camera. The high precision and accuracy of lidar have been vital to SDF, but its high cost, weight, and power demands limit its use in certain contexts. In these cases, automotive mmWave radar presents a cost-effective, lightweight, and energy-efficient alternative, albeit with a significant compromise in spatial resolution and accuracy. This study examines radar as an alternative nuclear SDF sensor by simplifying to the case of 2-D mapping and creating radar occupancy grids refined by a lidar-trained Pix2Pix Generative Adversarial Network (GAN) to approach lidar grid accuracy. We describe the creation and assessment of a radar-based LAMP equipped with six mmWave radar chips and a NaI radiation detector for nuclear SDF, conducting a direct comparison of its mapping precision and radiation localization capabilities with respect to lidar-based SDF.

The contemporary telecommunications system heavily depends on extensively spread optical fiber networks, which form its fundamental basis. Mechanical forces stemming from diverse ambient vibrations, including human activities and seismic movements, induce strains in these fibers. As a consequence, the light passing through the fibers experiences phase shifts. Consequently, these phase shifts can be detected throughout the entire fiber, offering insights into the original vibration occurrences. This characteristic renders them exceptionally suitable for distributed seismic sensing.

The United States Air Force Academy has developed techniques to characterize and identify satellites since 2014 using a 16-inch telescope on campus, as well as off-campus telescopes which comprise the Falcon Telescope Network. USAFA is upgrading its Falcon telescopes with new dual filter wheels and cameras. This will require updated calibrations for the new equipment along with a new capability to combine hyperspectral and polarimetric observations. For polarization calibrations, highly accurate star fields are used. Emission and absorption stars will be used to calibrate the hyperspectral observations. Once the calibrations are updated, the combined observations with the new dual filter wheels will be used to make observations of geosynchronous satellites. This effort focuses on these combined hyperspectral-polarimetric observations, characterizing differences observed between satellites, and development of techniques to exploit these differences for the identification or discrimination of satellites.

We experimentally and computationally investigated a proposed frequency-domain method for detecting and tracking cislunar spacecraft and near-earth asteroids using heliostat fields at night. The method oscillates the orientation of heliostats concentrating light from the stellar field and measures the light’s photocurrent power spectrum at sub-milliHertz resolution. Experimentally, we utilized one heliostat as Sandia’s National Solar Thermal Test Field and collected data from the night sky during the summer of 2023 for the purpose of evaluating background and noise. The data we collected indicate instrumentation responds to the stellar field and is somewhat consistent with our computational model. Computation indicates spacecraft or asteroids as dim as magnitude 18 moving at twice the sidereal rate can be detected using this method, and further improvement may be possible. A potential advantage of this frequency-domain method over imaging is that detectivity improves with apparent angular rate and number of heliostats.

The vision through scattering media is of great importance for many application fields such as driver assistance systems or medical diagnosis. By correlating the speckle pattern of two different illumination modes, we are able to retrieve object information through scattering media, as long as the two illumination wavefronts are within the memory effect range of the medium. This relaxes the requirements for the object’s position and dimension, which often exist in other memory effect protocols. For instance, with our method, we are able to extract information about objects embedded in multiply scattering media.

NIWC Atlantic and MZA Associates scientists proposed Bits-per-Joule Capacity (BPJC) as a standardized free space optical communications (FSOC) link performance metric at the Optica Imaging Congress in July 2024. The intent of this document is to motivate the establishment of FSOC performance standards through community research, discussion, and cooperation. Such standards currently do not exist, muddling the interpretation of FSOC link performance and unintentionally dissuading its widespread adoption. In this follow-up work, the authors seek to expand upon the original conference paper through application examples. This will include the development of a model, its limitations, and its applications to simple laboratory experiments.

Small angle scattering by relatively large atmospheric cloud/fog water droplets and ice crystals can cause significant contrast reduction and blurring of imagery. While this effect is quite well explained and verified in field experiments and sensor models, the extent to which aerosols, especially those of quite prevalent anthropogenic fine/ultra-fine/coarse mode play a role in image degradation remains to this date, a controversial topic. In this work, the contribution of aerosols to image blur will be revisited but with special focus on field data collected with a relatively large variety of ambient aerosol characterization and optical instrumentation. Ambient particulate/aerosol morphology and optical properties and trends will be correlated with collected multi-band imagery using instruments including nano-class condensation particle counters, a nephelometer to derive in real-time the composite aerosol phase functions, and a fog monitor to distinguish larger particulate (water droplet/ice crystal) contributions.

Atmospheric aerosols (fog) create degraded visual environments (DVEs), making it difficult to recover optical information from our surroundings. We have developed a low SWaP technique which characterizes these DVEs using an f-theta lens to capture the angular scattering profile of a pencil beam passed through an aerosol. These measurements are then compared to data taken in tandem by conventional DVE characterization techniques. We present this angular scattering measurement as a low SWaP alternative to current DVE characterization techniques to provide real-time data for implementation with signal recovery algorithms. SNL is managed and operated by NTESS under DOE NNSA contract DE-NA0003525.

This presentation will highlight recent advances by the Naval Information Warfare Center team in advancing homodyne interferometry for imaging applications to mitigate atmospheric turbulence in a passive imaging system. The presentation will showcase results from an outdoor imaging experiment using simultaneous data taken from a normal camera and an 18-aperture homodyne system. Performance between the 2 systems will be compared and quantified using Quick Response (QR) codes and barcode reading software.

NIWC Atlantic has continued to develop its approach to leveraging photon orbital angular momentum (OAM) eigenchannels to maximize photon delivery efficiency for free space optical applications. Free space laser beam propagation through atmospheric turbulence is improved through the identification and application of the time-dependent optimal OAM channel. However, OAM-enhanced range-finding or atmospheric sensing systems require some mechanism to determine optimal modality. Current implementations of rapidly tunable OAM technology use an electronic feedback system to identify the optimal transmission eigenchannel. This creates a limitation in long-range applications, where optical communications are typically poised to replace RF systems, but still require supporting communication channels to provide such electronic feedback. This paper discusses the methodology for determining optimal transmission eigenchannels, the approaches used to provide that feedback to the transmitter, and the results of atmospheric sensing experiments using OAM technology.

The hybrid wavefront sensor (HyWFS) integrates the ideal properties of both the Shack-Hartmann wavefront sensor (SHWFS) and the pyramid wavefront sensor (PyWFS) for adaptive optics. It enables an easy switch between the SHWFS and PyWFS algorithms in the absence of any moving parts and provides high-resolution detection. Calibration deviation may cause the wrong wavefront slope information calculation. This paper presents the performance of the HyWFS when the calibration of the pupil is changed. An experimental setup is established to perform closed-loop corrections with the position, size, and sampling of the pupil varying, respectively. The experimental results using both signal calculation algorithms exhibit the validity and robustness of the HyWFS in the actual AO system, providing reference data for astronomical adaptive optics.

In a recent paper, two approaches for identifying branch points from SHWFS measurements were explored using both simulated and experimental data. It was shown that these two approaches were complementary in their ability to detect branch points. Specifically, one approach worked better at detecting branch points located closer to the edge of the SHWFS’ lenslet pupils and the other approach worked better at detecting branch points located closer to the middle of the lenslet pupils. The synergistic nature of these two approaches yielded an overall improvement in branch-point identification compared to either approach alone. In this paper, we seek to leverage these prior findings to develop a branch-point tolerant phase reconstruction approach using SHWFS data. Here, a rotational phase field created using the branch-point identifications will be added to the least-squares phase estimated using the SHWFS data. The implications of this work for adaptive-optics compensation in high scintillation optical-turbulence environments will also be discussed.

In many fields, the problem of phase retrieval naturally arises in which one must recover phase information from only intensity measurements. One such field is in ground-based astronomy, where a system distorted by a point spread function (PSF) must be compensated for. This compensation must be in reference to the phase of the aberration distorting the system, requiring one to estimate the phase with access to only its PSF. To estimate the phase from a single PSF measurement, we find three simple components increase the probability of finding the correct solution: odd-number-sided polygons (odd polygons), basis representations, and a small neural network. These three components help to alleviate some of the non-convexity, reduce the dimensionality of the loss space, and allow for efficient parallel solutions to overcome remaining complexities. Simulation-based results are shown and comparisons to classical methods are made.

Explosive events create highly transient optical signatures that reveal a wealth of phenomenological information. Adequately characterizing such waveforms often requires high frame-rate detectors in the visible or infrared regions. These events are frequently categorized as point-source transient-events (PSTEs) due to their limited spatial extent with respect to the sensor spatial sampling. Traditional systems designed to characterize PSTEs face a performance trade-off between frame-rate and spatial format. High data rates, bandwidth, SWaP and detector cost must also be considered. To address these limitations, we present on the progress of designing and building a computational imaging system based on a rolling-shutter read-out combined with compressive-sensing reconstruction. This approach enables reconstruction of a PSTE signature at sampling rates on the order of the rolling-shutter rate compared to the system frame rate. We summarize results of our efforts and highlight the near real-time algorithms used for PSTE reconstruction.

In the present work, we propose a novel reference-less wavefront sensing method in a grating array-based wavefront sensor (GAWS). The diffracted +1 and -1 order spot arrays were utilized to eliminate the requirement of reference spots to measure the slope. The basic idea is that when there is a local tilt in the wavefront, the +1 and -1 diffracted spots move in opposite directions due to their optical phase conjugate relationship and share a common reference position. The proposed method facilitates estimation of the wavefront using a single camera frame and is advantageous in situations where a high-quality wavefront is not available as a reference. Preliminary simulation and experimental results are provided to corroborate our findings.

Timely detection of gas influx in a wellbore is necessary for preventing surface blowouts during drilling, which could result in loss of lives and equipment. Optical fiber-based distributed acoustic sensors (DAS) offer several advantages over traditional gauges for detecting gas influx by providing high-resolution spatiotemporal measurements along the entire length of the fiber-instrumented wellbore. However, computationally intensive processing of voluminous DAS datasets makes real-time decision-making using these sensors challenging. To address this, a novel computationally inexpensive approach for detecting gas in the wellbore is presented using Poincaré maps and Minnaert resonance analysis of DAS data. The methodology is demonstrated on real-time DAS data acquired during gas influx tests conducted in a 5163-ft-deep wellbore. The results demonstrate the effectiveness of the Poincaré plots for the timely detection of gas bubbles at different well depths by quantifying the chaos in the system and an improved understanding of gas bubble size distribution using the Minnaert frequency analysis.