SEARCH · Search NASA
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.
It is typical in edge and corner detection applications to examine a single scale or to consider some space of scales in the image without knowing which scale is appropriate for each location in the image.
Abstract We present the Swimmy (Subaru WIde-field Machine-learning anoMalY) survey program, a deep-learning-based search for unique sources using multicolored (grizy) imaging data from the Hyper Suprime-Cam Subaru Strategic Program (HSC-SSP). This program aims to detect unexpected, novel, and rare populations and phenomena, by utilizing the deep imaging data acquired from the wide-field coverage of the HSC-SSP. This article, as the first paper in the Swimmy series, describes an anomaly detection technique to select unique populations as “outliers” from the data-set. The model was tested with known extreme emission-line galaxies (XELGs) and quasars, which consequently confirmed that the proposed method successfully selected $\sim\!\! 60\%$–$70\%$ of the quasars and $60\%$ of the XELGs without labeled training data. In reference to the spectral information of local galaxies at z = 0.05–0.2 obtained from the Sloan Digital Sky Survey, we investigated the physical properties of the selected anomalies and compared them based on the significance of their outlier values. The results revealed that XELGs constitute notable fractions of the most anomalous galaxies, and certain galaxies manifest unique morphological features. In summary, deep anomaly detection is an effective tool that can search rare objects, and, ultimately, unknown unknowns with large data-sets. Further development of the proposed model and selection process can promote the practical applications required to achieve specific scientific goals.
The purpose of this project is the statistical evaluation of the occurrence of spectral lines in the gamma-ray burst spectra detected by the Ginga burst detector, and the comparison of the Ginga results to the BATSE observations. Two significant line features were detected in the Ginga bursts, but thus far none have been detected in the bursts BATSE detected. These line features may indicate the presence of strong magnetic fields in bursts, and therefore are important physical diagnostics of the conditions in the plasma which radiates the observed gamma-rays. The issue is whether there is a discrepancy between the Ginga and BATSE results; the potential discrepancy must be evaluated statistically. Even if BATSE line detections are announced, the statistical methodology we have developed can be used to estimate the rate at which different types of spectral features occur.
Clouds and aerosols play a significant role in determining the overall atmospheric radiation budget, yet remain a key uncertainty in understanding and predicting the future climate system. In addition to their impact on the Earth’s climate system, aerosols from volcanic eruptions, wildfires, man-made pollution events, and dust storms are hazardous to aviation safety and human health. Space-based lidar systems provide critical information about the vertical distributions of clouds and aerosols that greatly improve our understanding of the climate system. However, daytime data from backscatter lidars, such as the Cloud-Aerosol Transport System (CATS) on the International Space Station (ISS), must be averaged during science processing at the expense of spatial resolution to obtain sufficient signal-to-noise ratio (SNR) for accurately detecting atmospheric features. For example, 50% of all atmospheric features reported in daytime operational CATS data products require averaging to 60 km for detection. Furthermore, the single-wavelength nature of the CATS primary operation mode makes accurately typing these features challenging in complex scenes. This paper presents machine learning (ML) techniques that, when applied to CATS data, enable detection of atmospheric features during daytime operations with a horizontal resolution of 5 km compared to the 60 km horizontal resolution often required for daytime CATS data. A Convolutional Neural Network (CNN) trained using CATS standard data products also demonstrated the potential for improved cloud-aerosol discrimination compared to the operational CATS algorithms for cloud edges and complex near-surface scenes during daytime.
With the increasing concerns over climate change and carbon emissions, fault detection and diagnostics (FDD) of low–global warming potential (GWP) refrigerant supermarket refrigeration systems has gained great attention from academic and industrial sectors. Various FDD approaches have been developed to detect, identify, and diagnose faults to save energy, improve food quality, and protect the environment. Here, to mitigate the difficulty of collecting high-quality steady-state operational data in field operations faced by most model-based FDD methods, this study developed dynamic models of a low–GWP refrigerant (CO 2 ) supermarket refrigeration system. The model accuracy was validated using manufacturer data and experimental data. Simulations were conducted to predict the system dynamic response under two common operational faults—evaporator air path blockage fault and the display case door open fault—to identify fault patterns and define key dynamic behavior indexes for supporting FDD algorithm development.
This study proposes a novel method for signal detection and feature extraction based on the spectral correlation function, enabling improved characterization of grid-signal distortions. Our approach differs from existing treatments of signal distortion in its analysis of the varied spectral content of signals observed in real-world scenarios. The method we propose has state-of-the-art discriminative power that provides meaningful and understandable characterizations of various grid events and anomalies. In conclusion, to validate the approach, we use real world data from the Grid Event Signature Library, which is maintained jointly by Oak Ridge National Laboratory and Lawrence Livermore National Laboratory.
Explore the source record for details and available documents.
The electric grid is becoming increasingly cyber-physical with the addition of smart technologies, new communication interfaces, and automated grid-support functions. Because of this, it is no longer sufficient to only study the physical system dynamics, but the cyber system must also be monitored as well to examine cyber-physical interactions and effects on the overall system. To address this gap for both operational and security needs, cyber-physical situational awareness is needed to monitor the system to detect any faults or malicious activity. Techniques and models to understand the physical system (the power system operation) exist, but methods to study the cyber system are needed, which can assist in understanding how the network traffic and changes to network conditions affect applications such as data analysis, intrusion detection systems (IDS), and anomaly detection. In this paper, we examine and develop models of data flows in communication networks of cyber-physical systems (CPSs) and explore how network calculus can be utilized to develop those models for CPSs, with a focus on anomaly and intrusion detection. This provides a foundation for methods to examine how changes to behavior in the CPS can be modeled and for investigating cyber effects in CPSs in anomaly detection applications.
A presentation is given of 8.0-13.0 micron spectra (Delta lambda/lambda = 0.02-0.03) for six main belt asteroids, which range from 58 to 220 km in diameter and sample the five principal taxonomic classes (C, S, M, R and E). Narrow, well-defined silicate emission features are present on two of the asteroids, the C-type 19 Fortuna and the M-type 21 Lutetia. No comparable emission features are observed on the S-types 11 Parthenope and 14 Irene, the R-type 349 Dembowska or the E-type 64 Angelina.
Analysis of Airborne Imaging Spectrometer (AIS) data acquired in Australia has revealed a number of operational problems. Horizontal striping in AIS imagery and spectral distortions due to order overlap were investigated. Horizontal striping, caused by grating position errors can be removed with little or no effect on spectral details. Order overlap remains a problem that seriously compromises identification of subtle mineral absorption features within AIS spectra. A spectrometric model of the AIS was developed to assist in identifying spurious spectral features, and will be used in efforts to restore the spectral integrity of the data.
The paper presents 1-20-micron photometry of P/Giacobini-Zinner obtained at the NASA Infrared Telescope Facility, during 1985 June-September (r = 1.57-1.03 AU). A broad, weak 10-micron silicate emission feature was detected on August 26.6; a similar weak emission feature could have been hidden in the broadband photometry on other dates. The total scattering and emitting cross section of dust in the inner coma was similar to that in other short-period comets, but a factor of 10 (r = 1.56 AU) to 100 (r = 1.03 AU) lower than the amount of dust in Comet Halley. The thermal emission continuum can be fit with models weighted toward either small or large absorbing grains. The dust production rate near perihelion was about 100,000 g/s (small-grain model) to about 1,000,000 g/s (large-grain model). The corresponding dust/gas mass ratio on August 26 was about 0.1-1. A silicate-rich heterogeneous grain model with an excess of large particles is compatible with the observed spectrum of Giacobini-Zinner on August 26. Thus, weak or absent silicate emission does not necessarily imply an absence of silicates in the dust, although the abundance of silicate particles not greater than 1 micron radius must have been lower than in Comet Halley.
We present soft gamma-ray observations by the Oriented Scintillation Spectrometer Experiment (OSSE) on the Compton Gamma Ray Observatory (GRO) of the transient X-ray binary pulsar A0535+26. The observations were made 1994 February 8-17, immediately prior to the peak of a giant outburst. The phase averaged spectrum is complex and cannot be described by a single-component model. We find that structure in the spectrum above 100 keV can best be modeled by an absorption feature near 110 keV, which we interepret as the signature of cyclotron resonant scattering. Because of OSSE's 45 keV threshold, we are unable to make a definitive statement on the presence of a 55 keV absorption line; however, we can conclude that if this line does exist, it must have a smaller optical depth than the line at 110 keV. A first harmonic (=fundamental) cyclotron resonance at 110 keV corresponds to a magnetic field strength at the surface of the neutron star of approximately 1 x 10(exp 13) G (approximately 5 x 10(exp 12) G if the first harmonic is at 55 keV).
We present infrared spectra of four embedded protostars in the 750-1230 cm-1 (13.3-8.1 microns) range. For NGC 7538 IRS 9, a new band is reported at 1110 cm-1 (9.01 microns, and several others may be present near 785, 820, 900, 1030, and 1075 cm-1 (12.7, 12.2, 11.1, 9.71, and 9.30 microns). The band 1110 cm-1 is attributed to frozen NH3. Its position and width imply that the NH3 is frozen in a polar, H2O-rich interstellar ice component. The NH3/H2O ice ratio inferred for NGC 7538 IRS 9 is 0.1, making NH3 as important a component as CH3OH and CO2 in the polar ices along this line of sight. At these concentrations, hydrogen bonding between the NH3 and H2O can account for much of the enigmatic low-frequency wing on the 3240 cm-1 (3.09 microns) H2O interstellar ice band. The strength of the implied NH3 deformation fundamental at 1624 cm-1 (6.158 microns) can also account for the absorption at this position reported by ISO.
There are numerous challenges in successfully implementing and interpreting planetary ground penetrating radar (GPR) measurements. Many are due to substantial uncertainties in the target ground parameters and the intervening medium (i.e., the ionosphere). These uncertainties generate a compelling need for meaningful quantitative simulation of the planetary GPR problem. An accurate numerical model would enable realistic numerical GPR simulations using parameter regimes much broader than are possible in laboratory or field experiments. Parameters such as source bandwidth and power, surface and subsurface features, and ionospheric profiles could be rapidly iterated to understand their impact on GPR performance and the reliable interpretation of GPR data.
High energy cosmic rays constantly bombard the lunar regolith, producing (via nuclear evaporation) secondary 'albedo' or 'splash' particles like protons and neutrons, some of which escape back to space. Lunar Prospector and the Lunar Reconnaissance Orbiter (LRO), have shown that the energy distribution of albedo neutrons is modulated by the elemental composition of the lunar regolith, and by ice deposits in permanently shadowed polar craters. Here we investigate an analogous phenomenon with high energy ((is) approximately 100 MeV) lunar albedo protons.
Clouds and aerosols play a significant role in determining the overall atmospheric radiation budget, yet remain a key uncertainty in understanding and predicting the future climate system. In addition to their impact on the Earth’s climate system, aerosols from volcanic eruptions, wildfires, man-made pollution events and dust storms are hazardous to aviation safety and human health. Space-based lidar systems provide critical information about the vertical distributions of clouds and aerosols that greatly improve our understanding of the climate system. However, daytime data from backscatter lidars, such as the Cloud-Aerosol Transport System (CATS) on the International Space Station (ISS), must be averaged during science processing at the expense of spatial resolution to obtain sufficient signal-to-noise ratio (SNR) for accurately detecting atmospheric features. For example, 50% of all atmospheric features reported in daytime operational CATS data products require averaging to 60 km for detection. Furthermore, the single-wavelength nature of the CATS primary operation mode makes accurately typing these features challenging in complex scenes. This paper presents machine learning (ML) techniques that, when applied to CATS data, (1) increased the 1064 nm SNR by 75%, (2) increased the number of layers detected (any resolution) by 30%, and (3) enabled detection of 40% more atmospheric features during daytime operations at a horizontal resolution of 5 km compared to the 60 km horizontal resolution often required for daytime CATS operational data products. A Convolutional Neural Network (CNN) trained using CATS standard data products also demonstrated the potential for improved cloud-aerosol discrimination compared to the operational CATS algorithms for cloud edges and complex near-surface scenes during daytime.
Spectroscopy of WL 16 in the 8-13-micron region reveals a spectrum dominated by the hydrocarbon emission features at 7.7, 8.6, 11.25, and 12.7 microns. The emission plateau between the latter two features is detected, the first such detection in a young stellar object. The expected silicate feature is badly masked by the strong emission bands. Combining our spectra with the infrared spectral energy distribution indicates that silicate absorption is probably present.
We report on our effort to develop algorithms to detect, size, and classify geological features in planetary and asteroidal data sets.