Search NASA⌕ Search

SEARCH · Search NASA

Results for “Synthetic Image Generation”

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.

At least 109 records · Page 6

Building Maps for Terrain Relative Navigation Using Blender: An Open-Source Approach

A persistent challenge for vision-based navigation systems that compare imagery to a reference map is generating high quality maps with similar lighting conditions. Image rendering software can be used to apply variable lighting to reference maps or to generate synthetic imagery for test trajectories. While many image rendering software packages are available, with several developed specifically for spaceflight applications, there are often limitations due to cost, image fidelity, or flexibility. In this paper, we demonstrate the use of an open-source image rendering software, Blender, for use in Terrain Relative Navigation (TRN) applications. A scene in Blender was generated based on elevation data and satellite imagery of the region of West Texas used by Blue Origin for the operation of their New Shepard suborbital rocket. The Blender scene was validated by reproducing imagery collected during a flight of New Shepard in October 2020 and was further used to generate reference maps for use by a TRN algorithm on a subsequent New Shepard flight in August 2021. The work was performed under the NASA Safe and Precise Landing Integrated Capabilities Evolution (SPLICE) project, which is focused on technology advancement for precision landing and hazard avoidance. This work aims to lower the cost of entry and generally promote the adoption and advancement of vision-based navigation technologies.

Kyle W Smith↗

Computer Assembles Mosaics Of Satellite-SAR Imagery

Digital image-processing system assembles maplike mosaic images of large areas of Earth from overlapping images of smaller areas generated by spaceborne synthetic-aperture radar (SAR). Automatically registers and blends smaller images with each other. Eliminates need for time-consuming, tedious manual processing. Modifiable to accept images from spaceborne optical sensors.

Kwok, R.↗

Improving Sim-to-Real Transfer in Vision-Based Robot Navigation Via Instance-Level GAN-Based Data Augmentation

Achieving robust vision-based robotic tasks requires large amounts of data, which are often difficult to obtain in real-world scenarios. Simulators and synthetic data offer a cost-effective alternative, but the visual gap between simulation and reality hinders the performance of models when deployed in real-world environments. In this paper, we present a data augmentation pipeline that integrates a foundation model (Segment Anything Model) with an unsupervised image-to-image translation model (CycleGAN) for instance-level domain transfer from simulation to reality. This pipeline enables the generation of realistic labeled data from synthetic images for training supervised machine learning models in vision-based navigation tasks. We evaluate our approach on real-world data for ego-vehicle pose estimation, a critical autonomous navigation task involving the prediction of cross-track position and heading angle relative to road center line markings. The results of our tests show that our GAN-based data augmentation pipeline significantly outperforms models trained solely on simulation data or on data processed with standard image augmentation methods for sim-to-real transfer, enhancing model robustness and generalizability in real-world scenarios. Our method provides a scalable and flexible data augmentation tool for leveraging large synthetic datasets to enhance vision-based robotic navigation tasks.

artificial intelligence↗

A Feasibility Study of 4-D Tomography of Soft X-ray Magnetosheath Emissivities Using Multi-Spacecraft Measurements

Upcoming heliophysics missions utilize state-of-the-art wide field-of-view (FOV) imaging technology to measure and investigate the space plasma environment on a global scale. At Earth, remote sensing of soft X-ray emissions, which are generated via the charge exchange interaction between heavy solar wind ions and exospheric neutral atoms, is a promising means to investigate the global magnetosheath structure, its response to varying solar wind conditions, and the spatiotemporal properties of the dayside magnetic reconnection. Data analysis techniques such as optical tomography can provide additional structural and time-varying information from the observed target and thus enhance the mission’s scientific return. In this work, we simulate multiple and simultaneous observations of the dayside magnetosphere using soft X-ray imagers located at long-distance vantage points to reconstruct the time-dependent, three-dimensional (3-D) structure of the magnetosheath using a dynamic tomographic approach. The OpenGCCM MHD model is used to simulate the time-varying response of the magnetosheath to solar wind conditions and, subsequently, generate synthetic soft X-ray images from multiple spacecraft vantage points separated along a common orbit. A detailed analysis is then performed to identify the nominal set of spacecraft that produces the highest fidelity tomographic reconstruction of the magnetopause. This work aims to (i) demonstrate, for the first time, the use of dynamic tomography to retrieve the time-varying magnetosheath structure and (ii) identify a nominal mission design for multi-spacecraft configurations aiming for optical tomography.

3-D/4_d tomography↗

Demonstration of sub-micron UCN position resolution using room-temperature CMOS sensor

High spatial resolution of ultracold neutron (UCN) measurement is of growing interest to UCN experiments such as UCN spectrometers, UCN polarimeters, quantum physics of UCNs, and quantum gravity. Here we utilize physics informed deep learning to enhance the experimental position resolution and to demonstrate sub-micron spatial resolutions for UCN position measurements obtained using a room-temperature CMOS sensor, extending our previous work that demonstrated a position uncertainty of 1.5 microns. We explore the use of the open-source software Allpix Squared to generate experiment-like synthetic hit images with ground-truth position labels. We use physics-informed deep learning by training a fully connected neural network (FCNN) to learn a mapping from input hit images to output hit position. The automated analysis for sub-micron position resolution in UCN detection combined with the fast data rates of current and next generation UCN sources will enable improved precision for future UCN research and applications.

10B nanometer thin film↗

SU-D-207A-01: Female Pelvic Synthetic CT Generation Based On Joint Shape and Intensity Analysis

Purpose: To develop a method for generating female pelvic synthetic CT (MRCT) images from a single MR scan and evaluate its utility in radiotherapy. Methods: Under IRB-approval, an imaging sequence (T1-VIBE-Dixon) was acquired for 10 patients. This sequence yields 3 useful image volumes of different contrast (“in-phase” T1-weighted, fat and water). A previously published pelvic bone shape model was used to generate a rough bone mask for each patient. A modified fuzzy c-means classification was performed on the multi spectral MR data, with a regularization term that utilizes the prior knowledge provided by the bone mask and addresses the intensity overlap between different tissue types. A weighted sum of classification probabilities with attenuation values yielded MRCT volumes. The mean absolute error (MAE) between MRCT and real CT on various regions was calculated following deformable alignment (Velocity). Intensity modulated Treatment plans based on actual CT and MRCT were made and compared. Results: The average/standard deviation of MAE across 10 patients was 10.1/6.7 HU for muscle, 6.7/4.6 HU for fat, 136.9/53.5 HU for bony tissues under 850 HU (97% of total bone volume), 188.9/119.3 HU for bony tissues above 850 HU and 17.3/13.3 HU for intrapelvic soft tissues. Calculated doses were comparable for plans generated on CT and calculated using MRCT densities or vice versa, with differences in PTV D99% (mean/σ) of (–0.1/0.2 Gy) and (0.3/0.2 Gy), PTV D0.5cc of (–0.3/0.2 Gy) and (–0.4/1.7 Gy). OAR differences were similarly small for comparable structures, with differences in bowel V50Gy of (–0.3/0.2%) and (0.0/0.2%), femur V30Gy of (0.7/1.2%) and (0.2/1.2%), sacrum V20GY of (0.0/0.1%) and (–0.1/1.1%) and mean pelvic V20Gy of (0.0/0.1%) and (0.6/1.8%). Conclusion: MRCT based on a single imaging sequence in the female pelvis is feasible, with acceptably small variations in attenuation estimates and calculated doses to target and critical organs. Work supported by NIHR01EB016079.

60 APPLIED LIFE SCIENCES↗

Rotation measure synthesis applied to synthetic SKA images of galaxy clusters

Future observations with next-generation radio telescopes will help us to understand the presence and evolution of magnetic fields in galaxy clusters through determination of the so-called rotation measure (RM). In this work, we applied the RM synthesis technique to synthetic first phase Square Kilometre Array mid frequency element (i.e. the SKA1-MID) radio images of a pair of merging galaxy clusters, measured between 950 and 1750 MHz with a resolution of 10 arcsec and thermal noise of 0.1μJy beam -1 . The results of our RM synthesis analysis are compared with the simulation input parameters. We study two cases: one with radio haloes at the cluster centres and another without. Additionally, we found that the information obtained with RM synthesis is in general agreement with the input information; however, some discrepancies are present. We characterize them in this work, with the final goal of determining the potential impact of SKA1-MID on the study of cluster magnetic fields.

79 ASTRONOMY AND ASTROPHYSICS↗

Random Walks With Tweedie: A Unified View of Score-Based Diffusion Models [In the Spotlight]

We present a concise derivation for several influential score-based diffusion models that relies on only a few textbook results. Diffusion models have recently emerged as powerful tools for generating realistic, synthetic signals—particularly natural images—and often play a role in state-of-the-art algorithms for inverse problems in image processing. While these algorithms are often surprisingly simple, the theory behind them is not, and multiple complex theoretical justifications exist in the literature. Here, in this study, we provide a simple and largely self-contained theoretical justification for score-based diffusion models that is targeted towards the signal processing community. This approach leads to generic algorithmic templates for training and generating samples with diffusion models. We show that several influential diffusion models correspond to particular choices within these templates and demonstrate that alternative, more straightforward algorithmic choices can provide comparable results. This approach has the added benefit of enabling conditional sampling without any likelihood approximation.

97 MATHEMATICS AND COMPUTING↗

Software for Generating Troposphere Corrections for InSAR Using GPS and Weather Model Data

Atmospheric errors due to the troposphere are a limiting error source for spaceborne interferometric synthetic aperture radar (InSAR) imaging. This software generates tropospheric delay maps that can be used to correct atmospheric artifacts in InSAR data. The software automatically acquires all needed GPS (Global Positioning System), weather, and Digital Elevation Map data, and generates a tropospheric correction map using a novel algorithm for combining GPS and weather information while accounting for terrain. Existing JPL software was prototypical in nature, required a MATLAB license, required additional steps to acquire and ingest needed GPS and weather data, and did not account for topography in interpolation. Previous software did not achieve a level of automation suitable for integration in a Web portal. This software overcomes these issues. GPS estimates of tropospheric delay are a source of corrections that can be used to form correction maps to be applied to InSAR data, but the spacing of GPS stations is insufficient to remove short-wavelength tropospheric artifacts. This software combines interpolated GPS delay with weather model precipitable water vapor (PWV) and a digital elevation model to account for terrain, increasing the spatial resolution of the tropospheric correction maps and thus removing short wavelength tropospheric artifacts to a greater extent. It will be integrated into a Web portal request system, allowing use in a future L-band SAR Earth radar mission data system. This will be a significant contribution to its technology readiness, building on existing investments in in situ space geodetic networks, and improving timeliness, quality, and science value of the collected data

Moore, Angelyn W.↗

Finding the magnetopause location using soft X-ray observations and a statistical inverse method

Variability in the location and shape of the dayside magnetopause is attributed to magnetic reconnection, a fundamental process that enables the transfer of mass, energy, and momentum from the solar wind into the magnetosphere. The spatial and temporal properties of the magnetopause, under varying solar and magnetospheric conditions, remain largely unknown because empirical studies using in-situ observations are challenging to interpret. Global wide field-of-view (FOV) imaging is the only means to simultaneously observe the spatial distribution of the plasma properties over the vast dayside magnetospheric region and, subsequently, quantify the energy transport from the interplanetary medium into the terrestrial magnetosphere. Two upcoming missions, ESA/CAS SMILE and NASA’s LEXI will provide wide-field imagery of the dayside magnetosheath in soft X-rays, an emission generated by charge exchange interactions between high charge-state heavy ions of solar wind origin and exospheric neutral atoms. High-cadence two-dimensional observations of the magnetosheath will allow the estimation of dynamic properties of its inner boundary, the magnetopause, and enable studies of its response to changes in the solar wind dynamic pressure and interplanetary magnetic field orientation. This work introduces a statistically-based estimation approach based on inverse theory to estimate the spatial distribution of magnetosheath soft X-ray emissivities and, with this, identify the location of the magnetopause over the Sun−Earth line. To do so, we simulate the magnetosheath structure using the MHD-based OpenGGCM model and generate synthetic soft X-ray images using LEXI’s orbit and attitude information. Our results show that 3-D estimations using the described statistically-based technique are robust against Poisson-distributed shot noise inherent to soft X-ray images. Also, our proposed methodology shows that the accuracy of both three-dimensional (3-D) estimation and the magnetopause standoff distance calculation highly depends on the observational point.

Gonzalo Cucho-Padin↗

Digital enhancement of computerized axial tomograms

A systematic evaluation was conducted of certain digital image enhancement techniques performed in image space. Three types of images were used, computer generated phantoms, tomograms of a synthetic phantom, and axial tomograms of human anatomy containing images of lesions, artificially introduced into the tomograms. Several types of smoothing, sharpening, and histogram modification were explored. It was concluded that the most useful enhancement techniques are a selective smoothing of singular picture elements, combined with contrast manipulation. The most useful tool in applying these techniques is the gray-scale histogram.

Roberts, E., Jr.↗

TU-AB-BRA-02: An Efficient Atlas-Based Synthetic CT Generation Method

Purpose: A major obstacle for MR-only radiotherapy is the need to generate an accurate synthetic CT (sCT) from MR image(s) of a patient for the purposes of dose calculation and DRR generation. We propose here an accurate and efficient atlas-based sCT generation method, which has a computation speed largely independent of the number of atlases used. Methods: Atlas-based sCT generation requires a set of atlases with co-registered CT and MR images. Unlike existing methods that align each atlas to the new patient independently, we first create an average atlas and pre-align every atlas to the average atlas space. When a new patient arrives, we compute only one deformable image registration to align the patient MR image to the average atlas, which indirectly aligns the patient to all pre-aligned atlases. A patch-based non-local weighted fusion is performed in the average atlas space to generate the sCT for the patient, which is then warped back to the original patient space. We further adapt a PatchMatch algorithm that can quickly find top matches between patches of the patient image and all atlas images, which makes the patch fusion step also independent of the number of atlases used. Results: Nineteen brain tumour patients with both CT and T1-weighted MR images are used as testing data and a leave-one-out validation is performed. Each sCT generated is compared against the original CT image of the same patient on a voxel-by-voxel basis. The proposed method produces a mean absolute error (MAE) of 98.6±26.9 HU overall. The accuracy is comparable with a conventional implementation scheme, but the computation time is reduced from over an hour to four minutes. Conclusion: An average atlas space patch fusion approach can produce highly accurate sCT estimations very efficiently. Further validation on dose computation accuracy and using a larger patient cohort is warranted. The author is a full time employee of Elekta, Inc.

60 APPLIED LIFE SCIENCES↗

Iterative multi-task learning and inference from seismic images

Seismic interpretation aims to extract quantitative and interpretable attributes from a seismic image produced using some migration method to inform characteristics of a subsurface reservoir or target of interest. Current paradigms for computing seismic attributes mostly rely on single-task algorithms. We develop an iterative, multi-task machine learning method to learn and infer multiple attributes from a seismic image. This method is composed of two stages: a multi-task inference stage and a multi-modal, multi-task refinement stage. The basic mechanism of this method is that we train a multi-task inference neural network (NN) to estimate a set of attributes, including a relative geological time (RGT), a denoised higher-resolution (DHR) seismic image, and multiple fault attributes (including probability, dip, and strike), from a low-resolution, noisy seismic image; then we input the inferred attributes to a multi-task refinement NN to enhance the raw inference results iteratively. The two multi-task NNs are trained separately based on synthetic seismic images and associated attributes generated by a geological modeling algorithm. The software we intend to release is a PyTorch implementation of this multi-task learning method for both 2D and 3D cases along with scripts to run the training/validation. The algorithm and software can be a useful tool for automatic seismic interpretation.

Gao, Kai↗

Synthetic Vision Systems - Operational Considerations Simulation Experiment

Synthetic vision is a computer-generated image of the external scene topography that is generated from aircraft attitude, high-precision navigation information, and data of the terrain, obstacles, cultural features, and other required flight information. A synthetic vision system (SVS) enhances this basic functionality with real-time integrity to ensure the validity of the databases, perform obstacle detection and independent navigation accuracy verification, and provide traffic surveillance. Over the last five years, NASA and its industry partners have developed and deployed SVS technologies for commercial, business, and general aviation aircraft which have been shown to provide significant improvements in terrain awareness and reductions in the potential for Controlled-Flight-Into-Terrain incidents/accidents compared to current generation cockpit technologies. It has been hypothesized that SVS displays can greatly improve the safety and operational flexibility of flight in Instrument Meteorological Conditions (IMC) to a level comparable to clear-day Visual Meteorological Conditions (VMC), regardless of actual weather conditions or time of day. An experiment was conducted to evaluate SVS and SVS-related technologies as well as the influence of where the information is provided to the pilot (e.g., on a Head-Up or Head-Down Display) for consideration in defining landing minima based upon aircraft and airport equipage. The "operational considerations" evaluated under this effort included reduced visibility, decision altitudes, and airport equipage requirements, such as approach lighting systems, for SVS-equipped aircraft. Subjective results from the present study suggest that synthetic vision imagery on both head-up and head-down displays may offer benefits in situation awareness; workload; and approach and landing performance in the visibility levels, approach lighting systems, and decision altitudes tested.

Kramer, Lynda J.↗

Target Optimal Aperture Selection

This document describes the computation of optimal pixels for planetary transit targets. The method we describe is based on that used in the End-to-End Model (ETEM), which simulates Kepler science output. An optimal aperture for a target is defined as the set of pixels which maximizes the signal-to-noise ratio (SNR) for that target. The optimal aperture for a target is determined by using catalog data to generate, for that target, a synthetic image with all background stars and signals and a second synthetic image with the target star's flux only. These images are incorporated into a noise model, from which the SNR of each pixel is computed. The pixels around a target are summed in an order that maximizes the SNR with each term. As dimmer and dimmer pixels contribute to this sum, the SNR reaches a maximum and further terms decrease the SNR. The set of pixels whose SNR sums to this maximum value defines the optimal aperture.

Bryson, Stephen↗

Simulation of Image Performance Characteristics of the Landsat Data Continuity Mission (LDCM) Thermal Infrared Sensor (TIRS)

The next Landsat satellite, which is scheduled for launch in early 2013, will carry two instruments: the Operational Land Imager (OLI) and the Thermal Infrared Sensor (TIRS). Significant design changes over previous Landsat instruments have been made to these sensors to potentially enhance the quality of Landsat image data. TIRS, which is the focus of this study, is a dual-band instrument that uses a push-broom style architecture to collect data. To help understand the impact of design trades during instrument build, an effort was initiated to model TIRS imagery. The Digital Imaging and Remote Sensing Image Generation (DIRSIG) tool was used to produce synthetic "on-orbit" TIRS data with detailed radiometric, geometric, and digital image characteristics. This work presents several studies that used DIRSIG simulated TIRS data to test the impact of engineering performance data on image quality in an effort to determine if the image data meet specifications or, in the event that they do not, to determine if the resulting image data are still acceptable.

Landsat↗

Spectral Synthesis for Geostationary Satellite-to-Satellite Translation

Earth-observing satellites carrying multispectral sensors are widely used to monitor the physical and biological states of the atmosphere, land, and oceans. These satellites have different vantage points above the Earth and different spectral imaging bands resulting in inconsistent imagery from one to another. This presents challenges in building downstream applications. What if we could generate synthetic bands for existing satellites from the union of all domains? We tackle the problem of generating synthetic spectral imagery for multispectral sensors as an unsupervised image-to-image translation problem modeled with a variational autoencoder (VAE) and generative adversarial network (GAN) architecture. Our approach introduces a novel shared spectral reconstruction loss to constrain the high-dimensional feature space of multispectral images. Simulated experiments performed by dropping one or more spectral bands show that cross-domain reconstruction outperforms measurements obtained from a second vantage point. Our proposed approach enables the synchronization of multispectral data and provides a basis for more homogeneous remote sensing datasets.

Geostationary satellites↗

Synthetic Data Generation Using Machine Learning

Robust machine learning techniques for image analysis require a substantial amount of data to yield confident results. In the nuclear domain, data scarcity is a substantial challenge because there are so few facilities worldwide. This research focuses on being able alleviate the data scarcity problem by generating synthetic data to bridge the gap between large and small datasets. This work achieves that goal using a Generative Adversarial Network (GAN) architectural approach, by training a model on real-world data and expands that small dataset through synthetic data amendments. Model performance is impacted by the size of the real-world dataset and the number of training epochs utilized. This means that 1) It is important to develop your GAN to be optimized with the specific data type, and 2) approaches taken when training the GAN should be specialized to encompass important aspects of the dataset that it is generating. By taking a step to improve dataset sizes in this way, the gap between models trained by parties with significant amount of data and those without access to large data, closes, allowing for robust analyses of satellite imagery for nuclear domain applications.

11 NUCLEAR FUEL CYCLE AND FUEL MATERIALS↗