SEARCH · Engineering Papers
Results for “Image reconstruction”
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
Time-resolved 3D imaging of two-phase fluid flow inside a steel fuel injector using synchrotron X-ray tomography
The multiphase flow inside a diesel injection nozzle is imaged using synchrotron X-rays from the Advanced Photon Source at Argonne National Laboratory. Through acquisitions performed at several viewing angles and subsequent tomographic reconstruction, in-situ 3D visualization is achieved for the first time inside a steel injector at engine-like operating conditions. The morphology of the internal flow reveals strong flow separation and vapor-filled cavities (cavitation), the degree of which correlates with the nozzle's asymmetric inlet corner profile. Micron-scale surface features, which are artifacts of manufacturing, are shown to influence the morphology of the resulting liquid-gas interface. The data obtained at 0.1 ms time resolution exposes transient flow features and the flow development timescales are shown to be correlated with in-situ imaging of the fuel injector's hydraulically-actuated valve (needle). As more than 98.5% of the X-ray photon flux is attenuated within the steel injector body itself, we are posed with a unique challenge for imaging the flow within. Time-resolved imaging under these low-light conditions is achieved by exploiting both the refractive and absorptive properties of X-ray photons. The data-processing strategy converted these images with a signal-to-noise ratio of similar to 10 into a meaningful dataset for understanding internal flow and cavitation in a nozzle of diameter 200 mu m enclosed within 1-2 millimeters of steel.
Fast Hyperspectral Neutron Tomography
Hyperspectral neutron computed tomography is a tomographic imaging technique in which thousands of wavelength-specific neutron radiographs are measured for each tomographic view. In conventional hyperspectral reconstruction, data from each neutron wavelength bin are reconstructed separately, which is extremely time-consuming. These reconstructions often suffer from poor quality due to low signal-to-noise ratios. Consequently, material decomposition based on these reconstructions tends to produce inaccurate estimates of the material spectra and erroneous volumetric material separation. In this paper, we present two novel algorithms for processing hyperspectral neutron data: fast hyperspectral reconstruction and fast material decomposition. Both algorithms rely on a subspace decomposition procedure that transforms hyperspectral views into low-dimensional projection views within an intermediate subspace, where tomographic reconstruction is performed. The use of subspace decomposition dramatically reduces reconstruction time while reducing both noise and reconstruction artifacts. We apply our algorithms to both simulated and measured neutron data and demonstrate that they reduce computation and improve the quality of the results relative to conventional methods.
The Berkeley EUV spectrometer for ORFEUS
A novel EUV spectrometer is presented for the ORFEUS-SPAS mission. It uses a set of four varied line-space spherical diffraction gratings to obtain high-resolution spectra of point sources at wavelengths between 390 and 1200 A. The spectra are recorded with two detector units, each containing curved-surface microchannel plates and a delay-line anode-readout system. An independent optical system detects the image of the source in the entrance aperture and tracks the source as it drifts during an observation, enabling a reconstruction of the spectra postflight. The overall system performance is discussed and illustrated by synthetic spectra.
Extended far-infrared emission and star formation in Seyfert galaxies
An investigation into the extended distribution of far-infrared (FIR) emission associated with nearby Seyfert galaxies is made using a set of MEM reconstructions of IRAS Chopped Photometric Channel (CPC) data (Marston 1993). The data is compared to a set of HII/starburst galaxy images similarly processed in order to compare distributions and FIR color properties. It is shown that the central 1 kpc or so of Seyfert galaxies show extended FIR emission. FIR colors suggest that the bulk of this emission is not directly associated with an active nucleus. They further suggest that the origins of the majority of the emission is from heated dust associated with star formation surrounding the nucleus rather than dust heated by the active nucleus. Nearby Seyfert galaxies are shown to have a higher concentration of far-infrared emission from their centers than the HII/starburst galaxies and a number appear to reside in disk galaxies with relatively low ongoing star formation in their disks. An example of this is NGC 7582 which has a smooth disk but an active nucleus/starbust center.
Orbit Determination Strategy and Simulation Performance for OSIRIS-REx Proximity Operations
The Origins Spectral Interpretation Resource Identification Security Regolith Explorer (OSIRISREx)is a NASA New Frontiers mission to the near-earth asteroid Bennu that will rendez vousin 2018, create a comprehensive and detailed set of observations over several years, collect a regolith sample, and return the sample to Earth in 2023. The Orbit Determination (OD) team isa sub-section of the Flight Dynamics System responsible for generating precise reconstructions and predictions of the spacecraft trajectory. The OD team processes radiometric data, LIDAR, as well as center-finding and landmark-based Optical Navigation images throughout the proximity operations phase to estimate and predict the spacecraft location within several meters. Stringent knowledge requirements stress the OD teams concept of operations and procedures to produce verified and consistent high quality solutions for observation planning, maneuver planning, and onboard sequencing. This paper will provide insight into the OD concept of operations and summarize the OD performance expected during the approach and early proximity operation phases,based on our pre-encounter knowledge of Bennu. Strategies and methods used to compare and evaluate predicted and reconstructed solutions are detailed. The use of high fidelity operational tests during early 2017 will stress the teams concept of operations and ability to produce precise OD solutions with minimal turn-around delay.
Spectroscopic Fourth-order Coronagraph for the Characterization of Terrestrial Planets at Small Angular Separations from Host Stars
We propose a new approach for high-contrast imaging at the diffraction limit using segmented telescopes in a modest observation bandwidth. This concept, named “spectroscopic fourth-order coronagraphy,” is based on a fourth-order coronagraph with a focal-plane mask that modulates the complex amplitude of the Airy disk along one direction. While coronagraphs applying the complex amplitude mask can achieve the theoretical limit performance for any arbitrary pupils, the focal-plane mask severely limits the bandwidth. Here, focusing on the fact that the focal-plane mask modulates the complex amplitude along one direction, we noticed that the mask can be optimized for each spectral element generated by a spectrograph. We combine the fourth-order coronagraph with two spectrographs to produce a stellar spectrum on the focal plane and reconstruct a white pupil on the Lyot stop. Based on the wave-front analysis of an optical design applying an Offner-type imaging spectrograph, we found that the achievable contrast of this concept is 10{sup −10} at 1.2–1.5 times the diffraction limit over the wavelength range of 650–750 nm for the entrance pupil of the LUVOIR telescope. Thus, this coronagraph concept could bring new habitable planet candidates not only around G- and K-type stars beyond 20–30 pc but also around very nearby M-type stars. This approach potentially promotes the characterization of the atmospheres of nearby terrestrial planets with future on- and off-axis segmented large telescopes.
Object-space optimization of tomographic reconstructions for additive manufacturing
Volumetric 3D printing motivated by computed axial lithography enables rapid printing of homogeneous parts but requires a high dimensionality gradient-descent optimization to calculate image sets. Here we introduce a new, simpler approach to image-computation that algebraically optimizes a model of the printed object, significantly improving print accuracy of complex parts under imperfect material and optical precision by improving optical dose contrast between the target and surrounding regions. Quality metrics for volumetric printing are defined and shown to be significantly improved by the new algorithm. The approach is extended beyond binary printing to grayscale control of conversion to enable functionally graded materials. The flexibility of the technique is digitally demonstrated with realistic projector point spread functions, printing around occluding structures, printing with restricted angular range, and incorporation of materials chemistry such as inhibition. Finally, simulations show that the method facilitates new printing modalities such as printing into flat, rather than cylindrical packages to extend the applications of volumetric printing.
A large expert-curated cryo-EM image dataset for machine learning protein particle picking
Cryo-electron microscopy (cryo-EM) is a powerful technique for determining the structures of biological macromolecular complexes. Picking single-protein particles from cryo-EM micrographs is a crucial step in reconstructing protein structures. However, the widely used template-based particle picking process is labor-intensive and time-consuming. Though machine learning and artificial intelligence (AI) based particle picking can potentially automate the process, its development is hindered by lack of large, high-quality labelled training data. To address this bottleneck, we present CryoPPP, a large, diverse, expert-curated cryo-EM image dataset for protein particle picking and analysis. It consists of labelled cryo-EM micrographs (images) of 34 representative protein datasets selected from the Electron Microscopy Public Image Archive (EMPIAR). The dataset is 2.6 terabytes and includes 9,893 high-resolution micrographs with labelled protein particle coordinates. The labelling process was rigorously validated through 2D particle class validation and 3D density map validation with the gold standard. The dataset is expected to greatly facilitate the development of both AI and classical methods for automated cryo-EM protein particle picking.
PLIF Imaging of Capsule RCS Jets, Shear Layers, and Simulated Forebody Ablation
Planar laser-induced fluorescence (PLIF) has been used to investigate hypersonic flows associated with capsule reentry vehicles. These flows included reaction control system (RCS) jets, shear layer flow, and simulated forebody heatshield ablation. Pitch, roll, and yaw RCS jets were studied. PLIF obtained planar slices in these flowfields. These slices could be viewed individually or they could be combined using computer visualization techniques to reconstruct the three dimensional shape of the flow. The tests described herein were conducted in the 31-Inch Mach 10 Air Tunnel at NASA Langley Research Center. Improvements to many facets of the imaging system increased the efficiency and quality of both data acquisition, in addition to increasing the overall robustness of the system.
AIVT: Inference of turbulent thermal convection from measured 3D velocity data by physics-informed Kolmogorov-Arnold networks
We propose the artificial intelligence velocimetry-thermometry (AIVT) method to reconstruct a continuous and differentiable representation of the temperature and velocity in turbulent convection from measured three-dimensional (3D) velocity data. AIVT is based on physics-informed Kolmogorov-Arnold networks and trained by optimizing a loss function that minimizes residuals of the velocity data, boundary conditions, and governing equations. We apply AIVT to a set of simultaneously measured 3D temperature and velocity data of Rayleigh-Bénard convection, obtained by combining particle image thermometry and Lagrangian particle tracking. This enables us to directly compare machine learning results to true volumetric, simultaneous temperature and velocity measurements. We demonstrate that AIVT can reconstruct and infer continuous, instantaneous velocity and temperature fields and their gradients from sparse experimental data at a high resolution, providing an additional approach for understanding thermal turbulence.
3D-Stereoscopic Analysis of Solar Active Region Loops: SoHo/EIT Observations at Temperatures of 1.5-2.5 MK - 2
In this paper we study the three-dimensional (3D) structure of hot (T(sub e) approximately equals 1.5 - 2.5 MK) loops in solar active region NOAA 7986, observed on 1996 August 30 with the Extreme-ultraviolet Imaging Telescope (EIT) onboard the Solar and Heliospheric Observatory (SoHO). This complements a first study on cooler (T(sub e) approximately equals 1.0 - 1.5 MK) loops of the same active region, using the same method of Dynamic Stereoscopy to reconstruct the 3D geometry. We reconstruct the 3D-coordinates x(s), y(s), z(s), the density n(sub e)(s), and temperature profile T(sub e)(s) of 35 individual loop segments (as function of the loop coordinate s) using EIT 195 A and 284 A images. The major findings are: (1) All loops are found to be in hydrostatic equilibrium, in the entire temperature regime of T(sub e) = 1.0 - 2.5 MK; (2) The analyzed loops have a height of 2-3 scale heights, and thus only segments extending over about one vertical scale height have sufficient emission measure contrast for detection; (3) The temperature gradient over the lowest scale height is of order dT/ds is approximately 1 - 4 K/km; (4) The radiative loss rate is found to exceed the conductive loss rate by about two orders or magnitude, making thermal conduction negligible to explain the temperature structure of the loops; (5) A steady-state can only be achieved when the heating rate E(sub H) matches the radiative loss rate in hydrostatic equilibrium, requiring a heat deposition length lambda(sub H) of the half density scale height lambda, predicting a scaling law with the loop base pressure, EH varies as p(sub 0 exp 2). This favors coronal heating mechanisms that operate near the loop footpoints; (6) We find a reciprocal correlation between the loop pressure p(sub 0) and loop length L, i.e. p(sub 0) varies as 1/L, implying a scaling law of the steady-state requirement with loop length, i.e. E(sub H ) varies as 1/L(exp 2). The heating rate shows no correlation with the loop-aligned magnetic field component B(sub z) at the footpoints, but is correlated with the azimuthal field B(sub phi) = Bz(RDelta Phi/L) of a twisted loop, and is thus consistent with heating mechanisms based on field-aligned currents.
3D-Reconstruction of Tau Neutrinos in LArTPC Detectors
The Deep Underground Neutrino Experiment (DUNE) is a next-generation neutrino experiment currently under construction. DUNE will consist of two high-resolution neutrino interaction imaging detectors exposed to the world’s most intense neutrino beam, with the Near Detector at Fermilab and the Far Detector 1,300 km away in the Sanford Underground Research Facility in South Dakota, US. The high statistics and excellent resolution capabilities of DUNE's $^{40}$Ar detector will allow us to make precision studies of oscillation parameters capable of searching for CP violation in the lepton sector, testing interaction models, and studying phenomena that have until now, seemed too complex to measure, like $\nu_\tau$ detection and therefore, providing the completion of the 3-flavor neutrino paradigm. Knowledge of the $\nu_\tau$ detection can impact a broad spectrum of open questions. These include searching for non-standard neutrino interactions, constraining the unitarity of the PMNS matrix, searching for sterile neutrinos, and studying neutrino interactions. In the case of LArTPC data, the detector hits can be considered nodes in a graph, and the edges represent the spatial and temporal relationships between them. By using graph neural networks, it is possible to exploit these relationships and improve the accuracy of particle identification and reconstruction. During my presentation and specifically for tau neutrino reconstruction, I will show the effectiveness and reliability of our in-house developed graph neural network (GNN), NuGraph. This GNN classifies detector hits based on the particle type responsible for their production, assuring that the system accurately identifies and categorizes information based on its unique characteristics.
First Peek of ASTRO-H Soft X-Ray Telescope (SXT) In-Orbit Performance
ASTRO-H (Hitomi) is a Japanese X-ray astrophysics satellite just launched in February, 2016, from Tanegashima, Japan by a JAXA's H-IIA launch vehicle. It has two Soft X-ray Telescopes (SXTs), among other instruments, that were developed by the NASA Goddard Space Flight Center in collaboration with ISAS/JAXA and Nagoya University. One is for an X-ray micro-calorimeter instrument (Soft X-ray Spectrometer, SXS) and the other for an X-ray CCD camera (Soft X-ray Imager, SXI), both covering the X-ray energy band up to 15 keV. The two SXTs were fully characterized at the 30-m X-ray beam line at ISAS/JAXA. The combined SXT+SXS system effective area is about 250 and 300 cm(exp 2) at 1 and 6 keV, respectively, although observations were performed with the gate valve at the dewar entrance closed, which blocks most of low energy X-rays and some of high energy ones. The angular resolution for SXS is 1.2 arcmin (Half Power Diameter, HPD). The combined SXT+SXI system effective area is about 370 and 350 cm (exp 2) at 1 and 6 keV, respectively. The angular resolution for SXI is 1.3 arcmin (HPD). The both SXTs have a field of view of about 16 arcmin (FWHM of their vignetting functions).The SXT+SXS field of view is limited to 3 x 3 arcmin by the SXS array size. In-flight data available to the SXT team was limited at the time of this conference and a point-like source data is not available for the SXT+SXS. Although due to lack of attitude information we were unable to reconstruct a point spread function of SXT+SXI, according to RXJ1856.5-3754 data, the SXT seems to be working as expected in terms of imaging capability. As for the overall effective area response for both SXT+SXS and SXT+SXI, consistent spectral model fitting parameters with the previous measurements were obtained for Crab and G21.5-0.9 data. On the other hand, their 2-10 keV fluxes differ by about 20% at this point. Calibration work is still under progress. The SXT is the latest version of the aluminum foil X-ray mirror, which is extremely light-weight and very low cost, yet produces large effective area over a wide energy-band. Its area-mass ratio is the largest, 16 cm(exp 2)/kg, among ASTRO-H, Chandra, and XMM-Newton mirrors. The aluminum foil mirror is a still compelling technology depending on the mission science goal.
Applications Of Machine Learning to Gas Plume Analysis In Longwave Infrared Hyperspectral Images
Longwave infrared hyperspectral images can be used for gas plume analysis, as many gases exhibit distinct absorption features in this portion of the electromagnetic spectrum. In practice, accurately identifying weak gas signatures is difficult because the observed radiance is dominated by background radiance, which varies with material, temperature, and viewing conditions. Many gas plume analysis pipelines operate on single images, limiting the ability to leverage spatial and multi-view information that could enhance the analysis. The goal of this dissertation is to explore how machine learning and deep learning methods can complement classical approaches to improve gas plume identification in longwave infrared hyperspectral imagery, and to investigate the use of neural radiance fields for hyperspectral scene reconstruction.
Atomic‐Scale Surface Imaging of Bulk Epitaxial CsPbBr 3 Perovskite Single Crystals on Mica Using Light Assisted Scanning Tunneling Microscopy at Low‐Temperature (80 K)
Epitaxial single-crystalline CsPbBr 3 perovskite films on mica, prepared ex situ, are explored using a low-temperature scanning tunneling microscope (STM) by probing the unoccupied electronic states of their surface in ultra-high vacuum (UHV) at 80 K. Light-assisted STM measurements under a broadband illumination with visible light were employed to enhance and stabilize surface conductivity. STM imaging across the surface of macroscopic bulk CsPbBr 3 films reveals large flat terraces characterized by a specific type of surface reconstruction, consisting of parallel rows of U-shaped atomic nanostructures. These structures are spaced by 12 Å and exhibit an internal feature size of 5.1 Å. Density functional theory (DFT) calculations reproduce the experimental observations and reveal a competition between different orthorhombic CsPbBr 3 (110) surface reconstructions: a Cs-rich structure, identified as the most energetically stable, and three alternative Pb–Br-rich reconstructions, which are slightly higher in energy yet remain consistent with the STM data. Additional analyses that explicitly account for the mica substrate exclude the cubic CsPbBr 3 phase and other orthorhombic surface orientations, while showing that variations in the mica surface termination do not alter the preferred CsPbBr 3 (110) reconstruction. In conclusion, this combined approach thereby confirms our assignment and resolves previous STM interpretations of CsPbBr 3 .
Reconstructing porous media using generative flow networks
One area of intense scientific interest for the study of sandstones, carbonates, and shale at the pore scale is the use of limited image and petrophysical data to generate multiple realizations of a rock’s pore structure. Such images aid efforts to quantify uncertainty in petrophysical properties, including porosity–permeability transforms. We develop and evaluate a deep learning-based method to synthesize porous media volumes using a so-called generative flow model trained on x-ray microscope images of rock texture and pore structure. These models are optimized on a log-likelihood objective and they synthesize large and realistic three-dimensional images. Further, we demonstrate the rapid generation of sandstone image volumes that display realism as gauged by quantitative comparison of topological features using Minkowski functionals of porosity, specific surface area, and the Euler–Poincaré characteristic (i.e., pore connectivity). We also evaluate the single-phase permeability using Navier–Stokes and lattice Boltzmann methods and show that transport properties of the generated samples match measured trends.
A mesoscale 3D model of irradiated concrete informed via a 2.5 U-Net semantic segmentation
The concrete biological shield in light-water reactors is exposed to neutron and gamma irradiation, which deteriorates the concrete’s mechanical properties in the long term. To assess the irradiation-induced damage, predictive mechanical models are developed and used in parallel with the characterization of irradiated concrete samples. Realistic 3D simulation domains can drastically improve a model’s prediction. In this work, we utilized x-ray computed tomography (XCT) data of a concrete specimen to reconstruct its 3D microstructure. The XCT data shows low contrast between the concrete’s aggregates and cement paste, resulting in poor image segmentation when using traditional unsupervised techniques. To address this issue, we developed and trained a 2.5D U-Net model on only 24 pre-labeled XCT layers to segment 651 layers of the XCT data. The overall F1-score of the model is approximately 96%. Then, we created a 3D finite element (FE) mesh based on the stack of segmented images. The FE model contains radiation-induced expansion, damage, and creep. The constitutive equations are adapted to each phase (aggregates and cement paste). Here, we simulated the effects of neutron irradiation in the concrete specimen as well as the specimen’s mechanical response to uniaxial compression. Finally, model validation was performed using experimental data on similar concrete specimens in the literature.