SEARCH · Search NASA

Results for “Activity detection”

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.

SourceResearch connectionFrom yearThrough year

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 361 records · Page 20

All sources NASA NTRS DOE OSTI

DOE OSTI2025

Active Learning Meets Foundation Models: Fast Remote Sensing Data Annotation for Object Detection

Object detection in remote sensing demands extensive, high-quality annotations—a process that is both labor-intensive and time-consuming. In this work, we introduce a real-time active learning and semi-automated labeling framework that leverages foundation models to streamline dataset annotation for object detection in remote sensing imagery. For example, by integrating a Segment Anything Model (SAM), our approach generates mask-based bounding boxes that serve as the basis for dual sampling: (a) uncertainty estimation to pinpoint challenging samples, and (b) diversity assessment to ensure broad data coverage. Furthermore, our Dynamic Box Switching Module (DBS) addresses the well-known cold start problem for object detection models by replacing its suboptimal initial predictions with SAM-derived masks, thereby enhancing early-stage localization accuracy. Extensive evaluations on multiple remote sensing datasets plus a real-world user study, demonstrate that our framework not only reduces annotation effort, but also significantly boosts detection performance compared to traditional active learning sampling methods. The code for training and the user interface will be made available.

Burges, Marvin [ORNL] (ORCID:0000000312690769)↗

DOE OSTI2026

Developing a Deep Learning-Computer Vision Framework to Monitor Avian Interactions with Solar Energy Facility Infrastructure (Final Technical Report)

The project addressed an inability to monitor avian interactions with photovoltaic (PV) solar energy facilities necessary for understanding PV solar impacts on birds. In the project, machine-vision technology that continuously monitors avian activities at PV solar facilities was developed. The technology includes four machine-learning (ML) models, each of which accomplishes a specific task in detecting birds and classifying their activities in live or recorded videos—detecting and tracking moving objects, differentiating birds from other objects, detecting bird collisions with solar panels, and classifying non-collision bird activities around PV facilities. Major project outcomes include adoption by two of DOE SETO’s SolWEB projects, providing novel observational data on birds to promote co-location of PV solar development and habitat conservation, known as ecovoltaics.

14 SOLAR ENERGY↗

NASA NTRS2019

Ask-The-Expert: Minimizing Human Review for Big Data Analytics Through Active Learning

In this CIF project, we worked toward semi-automating knowledge discovery from anomaly detection algorithms through the use of active learning. Active learning is an area of research within machine learning that uses an "expert in the loop" to learn from large data sets that have very few annotations or labels available, and where providing such labels is expensive. In our case, the task can be defined as the identification of safety events from flight operational data. Since traditional anomaly detection algorithms cannot differentiate between operationally relevant and irrelevant statistical anomalies, Subject Matter Experts (SMEs) have a lengthy and expensive burden of investigating every example identified by the detection algorithm, classifying and labeling them as relevant or irrelevant. Active learningidentifies the unlabeled example for which a label would most improve the classifier, asks the domain expert for a label, and repeats this process until there are no more resources (time, budget) available for labeling or a minimum required performance is reached. A positive label indicates an operationally significant safety event whereas a negative label indicates otherwise. Based on these few labels we propose to build an active learning system that utilizes the SME's time in the most effective manner by iteratively asking for labels for as few informative instances as possible. Our work was proposed to be a stepping stone toward implementation and deployment of the system with user interface to be pursued by the Aviation Operations and Safety Program (AOSP) given its interest in safety monitoring and discovery of safety incidents.

aviation safety↗

NASA NTRS1991

Error latency measurements in symbolic architectures

Error latency, the time that elapses between the occurrence of an error and its detection, has a significant effect on reliability. In computer systems, failure rates can be elevated during a burst of system activity due to increased detection of latent errors. A hybrid monitoring environment is developed to measure the error latency distribution of errors occurring in main memory. The objective of this study is to develop a methodology for gauging the dependability of individual data categories within a real-time application. The hybrid monitoring technique is novel in that it selects and categorizes a specific subset of the available blocks of memory to monitor. The precise times of reads and writes are collected, so no actual faults need be injected. Unlike previous monitoring studies that rely on a periodic sampling approach or on statistical approximation, this new approach permits continuous monitoring of referencing activity and precise measurement of error latency.

Young, L. T.↗

NASA NTRS1999

Observational Studies of the Clearing Phase in Proto-Planetary Disks Surrounding Intermediate Mass Stars

A detailed study of circumstellar gas associated with young, intermediate-mass stars has demonstrated that, far from being unique or an infrequently occurring phenomenon, beta Pic-like infall activity is routinely observed in stars younger than 10-50 Myr when the observer's line of sight lies within 15 degrees of the disk mid-plane. Detailed studies of 2 Herbig Ae/Be stars, AB Aur and HD 163296 demonstrate that enhanced infall episodes last 20-60 hours, comparable to the duration of similar episodes in beta Pictoris. The infall activity is consistent with detection of the comae of swarms of star-grazing bodies of asteroidal to cometary composition. Episodic fluctuations in the infall activity are clearly present by approximately 6 Myr, and may indicate the presence of massive planets within the disk. This study has therefore, directly contributed to NASA's Origins of Planetary Systems theme by identifying under what conditions extra-solar planetesimals can be remotely sensed, indicating that such bodies appear to be routinely detectable among young stars in the 1-10 Myr range, and suggesting that temporal studies of spectroscopic variability may provide a means of identifying those systems harboring massive planets. This study has resulted in 2 refereed review papers, 13 other refereed papers, and 17 conference papers.

Grady, Carol A.↗

DOE OSTI2024

Avian Activity Classification Using Recurrent Networks to Fuse Videos with Metadata on Imbalanced Datasets

Activity classification plays a crucial role in various real-life scenarios involving both humans and animals. There is an increasing need for precise activity classification focused on avian-solar interactions, as the usage of solar energy facilities, such as photovoltaic array power stations, has been observed to impact bird species richness, behavior, and activity. However, there has been no work to develop an automated system to monitor and classify these avian-solar interactions. All current methods rely on human observers, which is time and human resources costly and subject to errors related to searcher efficiency. With the recent success of Deep Learning models in activity classification problems, this paper develops a recurrent neural network-based model to automatically classify six avian activities around solar energy facilities. Our proposed model integrates critical feature engineering metadata with video frame data, enabling improved learning and more accurate activity classification. Furthermore, we address the challenge of data imbalance during training and demonstrate the efficacy of our model in detecting and classifying different activities within video tracks. Additionally, we analyze the saliency/backpropagation map of the trained proposed model and validate its decision-making rationale.

Avian activity classification; bidirectional LSTM;↗

DOE OSTI2025

Arsenic Accumulation in Microbial Biomass and the Interpretation of Signals of Early Arsenic‐Based Metabolisms

Carbonaceous particles that concentrate arsenic in microbialites as old as ~3.5 Ga are similar to As-rich organic globules in modern microbialites. The former particles have been interpreted as tracers of As cycling by early microbial metabolisms. However, it is unclear if arsenic accumulation is a consequence of biological activity or passive postmortem binding of arsenic by organic matter during diagenesis in volcanically influenced, As-rich environments. Here, we address this uncertainty by evaluating the concentrations, speciation, and detectability of As in active or heat-killed biofilms formed by cyanobacteria or anoxygenic photosynthetic microbes exposed to environmentally relevant concentrations of As(III) or As(V) (50 μM to 3 mM). The genomes or metagenomes of these biofilms contain genes involved in detoxifying or energy-yielding As metabolisms. Biomass accumulates As from the solution in a concentration-dependent manner and with a preference for oxidized As(V) over As(III). Autoclaved biomass accumulates As even more strongly than active biomass, likely because living biofilms actively detoxify As. Active biofilms oxidize and reduce As and accumulate both As(III) and As(V), whereas a small fraction of As(V) can be reduced in inactive biofilms that bind As during diagenesis. Arsenic enrichments in the biomass are detectable by X-ray based spectroscopy techniques (XRF, EPMA-WDS) that are commonly used to analyze geological materials. These findings enable the reconstruction of past active and passive interactions of microbial biomass with arsenic in fossilized microbial biofilms and microbialites from the early Earth.

Madrigal‐Trejo, David↗

NASA NTRS1993

TeV gamma rays from millisecond pulsars and the detectability of globular clusters

I investigate the detectability of TeV gamma rays from millisecond pulsars, assuming the energy source that powers this emission is the loss of rotational energy of the spinning neutron star. I show that although individual sources may be too weak to detect, the contribution of many (about 1000) active millisecond pulsars may lead to currently detectable TeV gamma-ray emissions from globular clusters, provided the average pulsar converts about 0.1 percent of its spin-down energy into TeV gamma rays. Future detections of TeV gamma rays from globular clusters (or more restrictive upper limits on the fluxes) will lead to better estimates of the number of active millisecond pulsars the clusters contain, and to a better understanding of the pulsar emission mechanism. I also derive flux estimates for the following sources: (1) the Crab Nebula, (2) PSR 19374-21, PSR 1953 4-29, and PSR 1855 4-09, and (3) nearby millisecond pulsars in the galactic disk.

Smith, I. A.↗

NASA NTRS2012

Detection Thresholds of Falling Snow from Satellite-Borne Active and Passive Sensors

Precipitation, including rain and snow, is a critical part of the Earth's energy and hydrology cycles. Precipitation impacts latent heating profiles locally while global circulation patterns distribute precipitation and energy from the equator to the poles. For the hydrological cycle, falling snow is a primary contributor in northern latitudes during the winter seasons. Falling snow is the source of snow pack accumulations that provide fresh water resources for many communities in the world. Furthermore, falling snow impacts society by causing transportation disruptions during severe snow events. In order to collect information on the complete global precipitation cycle, both liquid and frozen precipitation must be collected. The challenges of estimating falling snow from space still exist though progress is being made. These challenges include weak falling snow signatures with respect to background (surface, water vapor) signatures for passive sensors over land surfaces, unknowns about the spherical and non-spherical shapes of the snowflakes, their particle size distributions (PSDs) and how the assumptions about the unknowns impact observed brightness temperatures or radar reflectivities, differences in near surface snowfall and total column snow amounts, and limited ground truth to validate against. While these challenges remain, knowledge of their impact on expected retrieval results is an important key for understanding falling snow retrieval estimations. Since falling snow from space is the next precipitation measurement challenge from space, information must be determined in order to guide retrieval algorithm development for these current and future missions. This information includes thresholds of detection for various sensor channel configurations, snow event system characteristics, snowflake particle assumptions, and surface types. For example, can a lake effect snow system with low (approx 2.5 km) cloud tops having an ice water content (IWC) at the surface of 0.25 g / cubic m and dendrite snowflakes be detected? If this information is known, we can focus retrieval efforts on detectable storms and concentrate advances on achievable results. Here, the focus is to determine thresholds of detection for falling snow for various snow conditions over land and lake surfaces. The results rely on simulated Weather Research Forecasting (WRF) simulations of falling snow cases since simulations provide all the information to determine the measurements from space and the ground truth. Sensitivity analyses were performed to better ascertain the relationships between multifrequency microwave and millimeter-wave sensor observations and the falling snow/underlying field of view. In addition, thresholds of detection for various sensor channel configurations, snow event system characteristics, snowflake particle assumptions, and surface types were studied. Results will be presented for active radar at Ku, Ka, and W-band and for passive radiometer channels from 10 to 183 GHz.

Gail Skofronick-Jackson↗

NASA NTRS1993

Preliminary analysis of the JAPE ground vehicle test data with an artificial neural network classifier

Remotely sensing and classifying military vehicles in a battlefield environment have been the source of much research over the past 20 years. The ability to know where threat vehicles are located is an obvious advantage to military personnel. In the past active methods of ground vehicle detection such as radar have been used, but with the advancement of technology to locate these active sensors, passive sensors are preferred. Passive sensors detect acoustic emissions, seismic movement, electromagnetic radiation, etc., produced by the target and use this information to describe it. Deriving the mathematical models to classify vehicles in this manner has been, and is, quite complex and not always reliable. However, with the resurgence of artificial neural network (ANN) research in the past few years, developing models for this work may be a thing of the past. Preliminary results from an ANN analysis to the tank signatures recorded at the Joint Acoustic Propagation Experiment (JAPE) at the US Army White Sands Missile Range, NM, in July 1991, are presented.

Larsen, Nathan F.↗

DOE OSTI2025

Eddy covariance towers as sentinels of abnormal radioactive material releases

Ensuring accurate detection and attribution of abnormal releases of radioactive material is critical for protecting human health and safety. Most commonly, such detection is accomplished via active monitoring approaches involving the collection of physical samples. Further, this is labor intensive and limits the temporal and spatial resolution of any detected events to a relatively coarse level. As an alternative first step towards passive monitoring, we developed an approach using eddy flux tower data records to identify signals from a known abnormal release and quantify the extent to which that signal also occurs at other times in the data record. Through two case studies, one of which targeted the Fukushima nuclear disaster and the other targeting an abnormal release event at a radioisotope production facility in Fleurus, Belgium, we tested our approach and identified several potential heretofore unidentified abnormal events that were consistent with atmospheric circulation patterns and/or wind direction from known release sites. Because our approach is relatively simple and is resistant to systematic errors in the observational record, it has broad applicability beyond specific constituents and ecosystem types to identify a wide variety of limited-duration anomalies in flux tower data to ensure human health and industrial safety.

54 ENVIRONMENTAL SCIENCES↗

NASA NTRS2014

Real-Time, Single-Step Bioassay Using Nanoplasmonic Resonator With Ultra-High Sensitivity

A nanoplasmonic resonator (NPR) comprising a metallic nanodisk with alternating shielding layer(s), having a tagged biomolecule conjugated or tethered to the surface of the nanoplasmonic resonator for highly sensitive measurement of enzymatic activity. NPRs enhance Raman signals in a highly reproducible manner, enabling fast detection of protease and enzyme activity, such as Prostate Specific Antigen (paPSA), in real-time, at picomolar sensitivity levels. Experiments on extracellular fluid (ECF) from paPSA-positive cells demonstrate specific detection in a complex bio-fluid background in real-time single-step detection in very small sample volumes.

Zhang, Xiang↗

NASA NTRS1982

A system for measuring thermal activation energy levels in silicon by thermally stimulated capacitance

One method being used to determine energy level(s) and electrical activity of impurities in silicon is described. The method is called capacitance transient spectroscopy (CTS). It can be classified into three basic categories: the thermally stimulated capacitance method, the voltage-stimulated capacitance method, and the light-stimulated capacitance method; the first two categories are discussed. From the total change in capacitance and the time constant of the capacitance response, emission rates, energy levels, and trap concentrations can be determined. A major advantage of using CTS is its ability to detect the presence of electrically active impurities that are invisible to other techniques, such as Zeeman effect atomic absorption, and the ability to detect more than one electrically active impurity in a sample. Examples of detection of majority and minority carrier traps from gold donor and acceptor centers in silicon using the capacitance transient spectrometer are given to illustrate the method and its sensitivity.

Cockrum, R. H.↗

DOE OSTI2024

Metatranscriptomic analysis reveals dissimilarity in viral community activity between an ice-free and ice-covered winter in Lake Erie

Winter is a relatively under-studied season in freshwater ecology. The paucity of wintertime surveys has led to a lack of knowledge regarding microbial community activity during the winter in Lake Erie, a North American Great Lake. Viruses shape microbial communities and regulate biogeochemical cycles by acting as top-down controls, yet very few efforts have been made to examine active virus populations during the winter in Lake Erie. Furthermore, climate change-driven declines in seasonal ice cover have been shown to influence microbial community structure, but no studies have compared viral community activity between different ice cover conditions. We surveyed surface water metatranscriptomes for viral hallmark genes as a proxy for active virus populations and compared activity metrics between ice-covered and ice-free conditions from two sampled winters. Transcriptionally active viral communities were detected in both winters, spanning diverse phylogenetic clades of putative bacteriophage (Caudoviricetes), giant viruses (Nucleocytoviricota, or NCLDV), and RNA viruses (Orthornavirae). However, viral community activity metrics revealed pronounced differences between the ice-covered and ice-free winters. Viral community composition was distinct between winters and viral hallmark gene richness was reduced in the ice-covered relative to the ice-free conditions. In addition, the observed differences in viral communities correlated with microbial community activity metrics. Overall, these findings contribute to our understanding of the viral populations that are active during the winter in Lake Erie and suggest that viral community activity may be associated with ice cover extent.

59 BASIC BIOLOGICAL SCIENCES↗

NASA NTRS2019

Swift-XRT Follow-up of Gravitational-wave Triggers in the Second Advanced LIGO/Virgo Observing Run

The Neil Gehrels Swift Observatory carried out prompt searches for gravitational-wave (GW) events detected by the LIGO/Virgo Collaboration (LVC) during the second observing run ("O2"). Swift performed extensive tiling of eight LVC triggers, two of which had very low false-alarm rates (GW170814 and the epochal GW170817), indicating a high confidence of being astrophysical in origin; the latter was the first GW event to have an electromagnetic counterpart detected. In this paper we describe the follow-up performed during O2 and the results of our searches. No GW electromagnetic counterparts were detected; this result is expected, as GW170817 remained the only astrophysical event containing at least one neutron star after LVC's later retraction of some events. A number of X-ray sources were detected, with the majority of identified sources being active galactic nuclei. We discuss the detection rate of transient X-ray sources and their implications in the O2 tiling searches. Finally, we describe the lessons learned during O2 and how these are being used to improve the Swift follow-up of GW events. In particular, we simulate a population of gamma-ray burst afterglows to evaluate our source ranking system's ability to differentiate them from unrelated and uncataloged X-ray sources. We find that ≈60%–70% of afterglows whose jets are oriented toward Earth will be given high rank (i.e., "interesting" designation) by the completion of our second follow-up phase (assuming that their location in the sky was observed), but that this fraction can be increased to nearly 100% by performing a third follow-up observation of sources exhibiting fading behavior.

N. J. Klingler↗

NASA NTRS2020

Hardware Verification and Validation for a Navigation Sensor Software Model in Support of Flight Vehicle Performance Analysis

… or, “It’s in the details, how to make complicated software perform like complicated hardware.” In attempts to minimize development time and quickly build an operational vehicle, NASA’s Space Launch System (SLS) has had to be intentional about integrated testing. Constraints on budget and schedule have required balance between testing needs and the desire for an integrated flight vehicle as soon as possible. To provide key insights early in design and analysis cycles, a large amount of effort has shifted into maturing and validating models at the component level with integrated testing as a means to validate their integration. In terms of SLS Navigation, this, and the model-based design approach have pushed explicit requirements for sensor models to be validated against flight hardware to high precision. This paper covers the approach taken to verify and validate the models for the two key navigation sensors on the SLS vehicle, the Redundant Inertial Navigation Sensor and the Rate Gyro Assembly. These models are used in performance evaluation, fault detection, and operations development extensively. Using a mix of data from hardware vendor documentation and testing reports, limited in-house testing, and integration activities, these models were able to be validated against flight hardware at multiple levels, from the internal software design to statistical behavior at the raw sensor and integrated box levels. The high level of insight into the hardware elements is instrumental to support flight certification activities and building confidence in SLS Navigation capability. Focused testing enabled additional insight and validation that proved invaluable and the resulting insights were used to focus and mature models. Additionally, of having validated performance-based hardware models enables a wide breadth of activities including detailed fault detection studies and integration into future vehicle frameworks, such as an upper stage and provide a valuable asset to continued SLS analysis and design.

Evan J Anzalone↗

NASA NTRS2020

Hardware Verification and Validation for a Navigation Sensor Software Model in Support of Flight Vehicle Performance Analysis

Thomas Park↗

NASA NTRS1977

Study of stratospheric-ionospheric coupling during thunderstorms and tornadoes

A continuous-wave-spectrum high-frequency Doppler sounder array with three transmitters at each of three sites was used to observe the dynamics of the coupling of energy between the stratosphere and the ionosphere. During times of severe weather activity wavelike disturbances have been detected on ground-based ionospheric sounding records as perturbations in electron densities. Infrasonic waves with wave periods of 3-7 min and with horizontal phase velocities of 600-800 m/s were observed when there was thunderstorm activity; gravity waves with wave periods of 10-15 min and horizontal phase velocities of 100-200 m/s were detected when there was tornado activity. Both triangulations from the cross correlation functions of the Doppler records based on an assumption of no background wind shear and ray-tracing computations including an assumed background wind shear indicate that the waves originated in the vicinity of the thunderstorms and tornadoes. A comparison of the wavelengths of the infrasonic and gravity waves observed at ionospheric heights and those in cloud-top pictures from satellites show that they are all of the order of 100-300 km.

Hung, R. J.↗