Engineering Papers⌕ Search

SEARCH · Engineering Papers

Results for “Image Classification”

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 973 records · Page 54

Guatemala and Panama Urban Development: Evaluating the Effects of Urban Expansion on Social and Environmental Vulnerability in Guatemala and Panama

Central America is experiencing rapid and unregulated urban expansion, which is contributing to an increase in socioeconomic and environmental risks including inequities in infrastructure and housing accessibility, biodiversity loss, vulnerability to natural disasters, and negative health outcomes. NASA DEVELOP, in partnership with NASA SERVIR, Sistema de la Integración Centroamericana (SICA), Secretariat of Central American Social Integration (SISCA), Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ), and Centro de Coordinación para la Prevención de los Desastres en América Central y República Dominicana (CEPRENEDAC), examined changes in urban extent, characterized roofing material type, and analyzed vulnerability within urban areas in two Central American cities, Guatemala City and Panama City. The team used land cover imagery from Landsat 5 Thematic Mapper (TM), Landsat 7 Enhanced Thematic Mapper Plus (ETM+), Landsat 8 Operational Land Imager (OLI), and Landsat 9 OLI-2 to map urban extent, and surface reflectance data from Maxar Worldview to identify roofing material types. Socioeconomic and environmental data were used to assess vulnerability. Results depict how the two cities have expanded from 2000 to present day and highlight areas of greatest vulnerability within each urban area. The supervised classification of roofing materials performed well but could be improved with a few enhancements. Findings can help partner organizations improve monitoring of urbanization and inform their planning and decision-making while prioritizing disaster prevention, public health, and environmental integrity. Additionally, these case studies can be used to inform future, similar work elsewhere in Central America to aid in understanding urbanization and its associated challenges.

Jennifer Ruiz↗

Onboard Classification of Hyperspectral Data on the Earth Observing One Mission

Remote-sensed hyperspectral data represents significant challenges in downlink due to its large data volumes. This paper describes a research program designed to process hyperspectral data products onboard spacecraft to (a) reduce data downlink volumes and (b) decrease latency to provide key data products (often by enabling use of lower data rate communications systems). We describe efforts to develop onboard processing to study volcanoes, floods, and cryosphere, using the Hyperion hyperspectral imager and onboard processing for the Earth Observing One (EO-1) mission as well as preliminary work targeting the Hyperspectral Infrared Imager (HyspIRI) mission.

cryosphere↗

Lineament and polygon patterns on Europa

A classification scheme is presented for the lineaments and associated polygonal patterns observed on the surface of Europa, and the frequency distribution of the polygons is discussed in terms of the stress-relief fracturing of the surface. The lineaments are divided on the basis of albedo, morphology, orientation and characteristic geometry into eight groups based on Voyager 2 images taken at a best resolution of 4 km. The lineaments in turn define a system of polygons varying in size from small reticulate patterns the limit of resolution to 1,000,000 sq km individuals. Preliminary analysis of polygon side frequency distributions reveals a class of polygons with statistics similar to those found in complex terrestrial terrains, particularly in areas of well-oriented stresses, a class with similar statistics around the antijovian point, and a class with a distribution similar to those seen in terrestrial tensional fracture patterns. Speculations concerning the processes giving rise to the lineament patterns are presented.

Pieri, D. C.↗

Patterns of vegetation in the Owens Valley, California

Spectral characteristics of semi-arid shrub communities were examined using Airborne Imaging Spectrometer (AIS) data collected in the tree mode on 23 May 1985. Mesic sites with relatively high vegetation density and distinct zonation patterns exhibited greater spectral signature variations than sites with more xeric shrub communities. Spectral signature patterns were not directly related to vegetation density or physiognomy, although spatial maps derived from an 8-channel maximum likelihood classification were supported by photo-interpreted surface features. In AIS data, the principal detected effect of shrub vegetation on the alluvial fans is to lower reflectance across the spectrum. These results are similar to those reported during a period of minimal physiological activity in autumn, indicating that shadows cast by vegetation canopies are an important element of soil-vegetation interaction under conditions of relatively low canopy cover.

Ustin, S. L.↗

Satellite radar for monitoring forest resources

An evaluation is made of the computer analysis results of a study which used Seasat satellite radar data obtained in 1978 and Shuttle Imaging Radar-B data obtained in 1984. The change-detection procedures employed demonstrate that deforestation and reforestation activities can be effectively monitored on the basis of radar data gathered at satellite altitudes. The computer-processing techniques applied to the data encompassed (1) overlay display, (2) ratios, (3) differences, (4) principal-component analysis, and (5) classification; of these, overlay display is noted to quickly and easily yield a qualitative display of the multidate data.

Hoffer, Roger M.↗

The Chandra Multi-Wavelength Project (ChaMP): A Serendipitous X-Ray Survey Using Chandra Archival Data

The launch of the Chandra X-ray Observatory in July 2000 opened a new era in X-ray astronomy. Its unprecedented, < 1" spatial resolution and low background is providing views of the X-ray sky 10-100 times fainter than previously possible. We have begun to carry out a serendipitous survey of the X-ray sky using Chandra archival data to flux limits covering the range between those reached by current satellites and those of the small area Chandra deep surveys. We estimate the survey will cover about 8 sq.deg. per year to X-ray fluxes (2-10 keV) in the range 10(exp -13) - 6(exp -16) erg cm2/s and include about 3000 sources per year, roughly two thirds of which are expected to be active galactic nuclei (AGN). Optical imaging of the ChaMP fields is underway at NOAO and SAO telescopes using g',r',z' colors with which we will be able to classify the X-ray sources into object types and, in some cases, estimate their redshifts. We are also planning to obtain optical spectroscopy of a well-defined subset to allow confirmation of classification and redshift determination. All X-ray and optical results and supporting optical data will be place in the ChaMP archive within a year of the completion of our data analysis. Over the five years of Chandra operations, ChaMP will provide both a major resource for Chandra observers and a key research tool for the study of the cosmic X-ray background and the individual source populations which comprise it. ChaMP promises profoundly new science return on a number of key questions at the current frontier of many areas of astronomy including solving the spectral paradox by resolving the CXRB, locating and studying high redshift clusters and so constraining cosmological parameters, defining the true, possibly absorbed, population of quasars and studying coronal emission from late-type stars as their cores become fully convective. The current status and initial results from the ChaMP will be presented.

Wilkes, Belinda↗

Chapter 19A – Radar Cross Section

The material in this Section will provide the novice Flight Test Engineer (FTE) with a brief overview of the radar cross section (RCS) concept. Brief details of radar range descriptions, operations, target calibrations, and data reduction are included. It is noted that these range operations concentrate on the "static" test range rather than the much more sophisticated "dynamic" test range. The dynamic test range is what the FTE will probably work with. Radar reflectivity measurement has developed over the last two decades from a relatively simple endeavor involving the measurement of target RCS amplitude statistics to involving wide band, coherent systems that can measure high resolution images of targets as well as, in many cases, the polarization and phasing properties. This rapid growth in RCS technology has occurred because of the increased use of radar in today's commercial and military systems. In general, the goal for commercial systems is to enhance radar reflectivity, whereas the military goal is to reduce radar reflectivity. Also, classification and identification are important for military purposes because in adverse weather radar may be the only system that may be capable of separating enemy targets from friendly ones. This Section will discuss the fundamentals of RCS. Since the flight test techniques associated with RCS measurements are very similar to those of antenna pattern measurements, both will be presented at the same time in Section 19B Antenna Radiation Pattern Measurements.

Robert W Borek↗

Hawaiʻi Urban Development: Mapping Urban Infrastructure and Linkages to Wastewater Impacts across Hawaiʻi

Coastal ecosystems and human health across the Hawaiʻian islands are under threat from cesspools and other unpermitted wastewater infrastructure. Cesspools, or home sewage systems that hold wastewater effluent in porous concrete containers underground, discharge raw, untreated sewage directly into waterways. In fact, it is estimated that approximately 88,000 cesspools release 55 million gallons of raw sewage into Hawaiʻi’s waterways each day. Our partners, The Nature Conservancy, Hawaiʻi Regional Office (TNC) and University of Hawaiʻi, Mānoa (UHM), are heavily involved with the implementation and impact of Act 125, which was passed in 2017 by Hawaiʻi State Legislature to help combat these effects. This requires the conversion of cesspools to be upgraded and either converted to approved onsite disposal systems or connected to a sewer system so that the wastewater may be treated and to avoid leaching into the groundwater. Thus, identifying and locating cesspool and other unpermitted wastewater infrastructure is key to state and partner goals. Utilizing NASA Earth observations including Landsat 7 Enhanced Thematic Mapper Plus (ETM+), Landsat 8 Operational Land Imager (OLI), and Landsat 9 OLI-2 imagery, as well as pre-classified Sentinel-2 Multispectral Instrument (MSI) imagery in Esri’s Living Atlas, our team showcased the various perspectives that wastewater pollution research can use to assist in identifying and tracking urban development and concrete condition. This process was demonstrated by utilizing supervised object-based classification models, calculating various urban indices, and finding building size estimates to calculate effluent levels. Additionally, providing our partners with a GIS Methodology that can be built on to reach their future needs.

urban classification↗

The evaluation of alternate methodologies for land cover classification in an urbanizing area

The usefulness of LANDSAT in classifying land cover and in identifying and classifying land use change was investigated using an urbanizing area as the study area. The question of what was the best technique for classification was the primary focus of the study. The many computer-assisted techniques available to analyze LANDSAT data were evaluated. Techniques of statistical training (polygons from CRT, unsupervised clustering, polygons from digitizer and binary masks) were tested with minimum distance to the mean, maximum likelihood and canonical analysis with minimum distance to the mean classifiers. The twelve output images were compared to photointerpreted samples, ground verified samples and a current land use data base. Results indicate that for a reconnaissance inventory, the unsupervised training with canonical analysis-minimum distance classifier is the most efficient. If more detailed ground truth and ground verification is available, the polygons from the digitizer training with the canonical analysis minimum distance is more accurate.

Smekofski, R. M.↗

Automated cloud screening of AVHRR imagery using split-and-merge clustering

Previous methods to segment clouds from ocean in AVHRR imagery have shown varying degrees of success, with nighttime approaches being the most limited. An improved method of automatic image segmentation, the principal component transformation split-and-merge clustering (PCTSMC) algorithm, is presented and applied to cloud screening of both nighttime and daytime AVHRR data. The method combines spectral differencing, the principal component transformation, and split-and-merge clustering to sample objectively the natural classes in the data. This segmentation method is then augmented by supervised classification techniques to screen clouds from the imagery. Comparisons with other nighttime methods demonstrate its improved capability in this application. The sensitivity of the method to clustering parameters is presented; the results show that the method is insensitive to the split-and-merge thresholds.

Gallaudet, Timothy C.↗

Benchmark Models for Classification of Radiation Type Induced in Immune Cells

NASA Biological and Physical Sciences and the Science Mission Directorate have published a benchmark dataset of mouse immune cells subjected to radiation-induced DNA damage. The dataset comprises ML-ready microscopic imagery of said cells, including labels indicating radiation type and dose. The machine learning team at NASA Interagency Implementation and Advanced Concept Team (IMPACT) created multiple benchmark models. Initially, we conducted a preliminary analysis using thresholding. The algorithm used thresholds on average brightness of the available images to classify them into their respective radiation type. We also tested machine learning approaches. Convolutional Neural Networks (CNN) emerged as the best-performing model. This poster presents the benchmark scores obtained by the models.

Vishal Perekadan↗

Application of multispectral scanner data to the study of an abandoned surface coal mine

The utility of aircraft multispectral scanner data for describing the land cover features of an abandoned contour-mined coal mine is considered. The data were obtained with an 11 band multispectral scanner at an altitude of 1.2 kilometers. Supervised, maximum-likelihood statistical classifications of the data were made to establish land-cover classes and also to describe in more detail the barren surface features as they may pertain to the reclamation or restoration of the area. The scanner data for the surface-water areas were studied to establish the variability and range of the spectral signatures. Both day and night thermal images of the area are presented. The results of the study show that a high degree of statistical separation can be obtained from the multispectral scanner data for the various land-cover features.

Spisz, E. W.↗

Asian Dust Weather Categorization with Satellite and Surface Observations

This study categorizes various dust weather types by means of satellite remote sensing over central Asia. Airborne dust particles can be identified by satellite remote sensing because of the different optical properties exhibited by coarse and fine particles (i.e. varying particle sizes). If a correlation can be established between the retrieved aerosol optical properties and surface visibility, the intensity of dust weather can be more effectively and consistently discerned using satellite rather than surface observations. In this article, datasets consisting of collocated products from Moderate Resolution Imaging Spectroradiometer Aqua and surface measurements are analysed. The results indicate an exponential relationship between the surface visibility and the satellite-retrieved aerosol optical depth, which is subsequently used to categorize the dust weather. The satellite-derived spatial frequency distributions in the dust weather types are consistent with China s weather station reports during 2003, indicating that dust weather classification using satellite data is highly feasible. Although the period during the springtime from 2004 to 2007 may be not sufficient for statistical significance, our results reveal an increasing tendency in both intensity and frequency of dust weather over central Asia during this time period.

Lin, Tang-Huang↗

Exploration Clinical Decision Support System: Medical Data Architecture

The Exploration Clinical Decision Support (ECDS) System project is intended to enhance the Exploration Medical Capability (ExMC) Element for extended duration, deep-space mission planning in HRP. A major development guideline is the Risk of "Adverse Health Outcomes & Decrements in Performance due to Limitations of In-flight Medical Conditions". ECDS attempts to mitigate that Risk by providing crew-specific health information, actionable insight, crew guidance and advice based on computational algorithmic analysis. The availability of inflight health diagnostic computational methods has been identified as an essential capability for human exploration missions. Inflight electronic health data sources are often heterogeneous, and thus may be isolated or not examined as an aggregate whole. The ECDS System objective provides both a data architecture that collects and manages disparate health data, and an active knowledge system that analyzes health evidence to deliver case-specific advice. A single, cohesive space-ready decision support capability that considers all exploration clinical measurements is not commercially available at present. Hence, this Task is a newly coordinated development effort by which ECDS and its supporting data infrastructure will demonstrate the feasibility of intelligent data mining and predictive modeling as a biomedical diagnostic support mechanism on manned exploration missions. The initial step towards ground and flight demonstrations has been the research and development of both image and clinical text-based computer-aided patient diagnosis. Human anatomical images displaying abnormal/pathological features have been annotated using controlled terminology templates, marked-up, and then stored in compliance with the AIM standard. These images have been filtered and disease characterized based on machine learning of semantic and quantitative feature vectors. The next phase will evaluate disease treatment response via quantitative linear dimension biomarkers that enable image content-based retrieval and criteria assessment. In addition, a data mining engine (DME) is applied to cross-sectional adult surveys for predicting occurrence of renal calculi, ranked by statistical significance of demographics and specific food ingestion. In addition to this precursor space flight algorithm training, the DME will utilize a feature-engineering capability for unstructured clinical text classification health discovery. The ECDS backbone is a proposed multi-tier modular architecture providing data messaging protocols, storage, management and real-time patient data access. Technology demonstrations and success metrics will be finalized in FY16.

Biomedical support↗

A land use and environmental impact analysis of the Norfolk-Portsmouth SMSA

The feasibility of using remote sensing techniques for land use and environmental assessment in the Norfolk-Portsmouth area is discussed. Data cover the use of high altitude aircraft and satellite remote sensing data for: (1) identifying various heirarchial levels of land use, (2) monitoring land use changes for repetitive basis, (3) assessing the impact of competing land uses, and (4) identifying areas of potential environmental deterioration. High altitude aircraft photographs (scale 1:120,000) acquired in 1959, 1970, and 1972, plus Earth Resources Technology Satellite (ERTS-1) color composite images acquired in 1972 were used for the land use and environmental assessments. The high altitude aircraft photography, as expected, was successfully used to map Level 1, Level 2, as well as some urban Level 3 land use categories. However, the detail of land use analysis obtainable from the ERTS imagery exceeded the expectations for the U.S. Geological Survey's land use classification scheme. Study results are consistent with the initial investigation which determined Level 1 land use change to be 16.7 square km per year.

Mitchel, W. B.↗

Assessing Flooding Vulnerability to Assist High Water Intervention and Urban Planning Programs in the Charles River Watershed with NASA DEVELOP

The Charles River watershed intersects 35 municipalities within the Boston Metropolitan Area and has a total population of 1.2 million, making it one of the most densely populated watersheds in New England. In recent years, the watershed has observed higher rates of flood inundation, mainly due to increased development, extreme precipitation events, and increased surface runoff. As the frequency of flooding events increases and a changing climate poses an ongoing threat to local communities, governments and organizations in Massachusetts are in need of accurate flood risk assessments. NASA DEVELOP partnered with the Charles River Watershed Association, the Town of Natick’s Office of Sustainability, and the Massachusetts Audubon Society to assess flood vulnerability and susceptibility in the watershed. The team used Landsat 5 Thematic Mapper, Landsat 8 Operational Land Imager, Sentinel-1 C-Band Synthetic Aperture Radar, and Sentinel-2 MultiSpectral Instrument to assess the feasibility of identifying the extent of past flood events using remote sensing. After identifying images that overlapped with reported flood events, the team concluded that it was not feasible to use Earth observation data to detect localized flooding in the time available for this study. Instead, the Federal Emergency Management Agency (FEMA) 100-year floodplain was used as a proxy for areas where flooding may occur. The team used statistical regression analysis and validation and supervised classification to develop a flood susceptibility map, incorporating several flood conditioning factors. The susceptibility maps were calibrated to various thresholds, including two that highlight hypothetical flooding under more liberal and more conservative planning scenarios. These were overlaid with demographic and socioeconomic data to create flood vulnerability maps. The team’s flood susceptibility maps showed an improvement in capturing known flood events over the FEMA 100-year and 500-year floodplain maps. These results will be improved with the addition of stormwater drainage mapping and precipitation data. Results can be used to fill in the gaps to help the stakeholders understand their communities’ vulnerability and susceptibility to flooding and improve their preparedness plans.

Trista Brophy↗

Using X-Ray Computed Tomography to Catalog Rock Fragments in Apollo Drive Tube 73002

Overview: The Apollo missions collected 382 kg of rock, regolith, and core samples from six locations on the nearside of the Moon. Approximately 84% by mass of the Apollo collection remains in pristine condition within the curation facility at Johnson Space Center (i.e., never allocated, continuously stored in dry-N2 purged cabinets, exposure history restricted to Teflon, stainless steel, and Al-metal). Although most Apollo samples have been well characterized, there are several types of samples that have remained wholly or largely unstudied since their return, and/or that have been cu-rated under special conditions, e.g., frozen samples, samples stored in He-purged environment, and previously unopened drive tubes. NASA solicited proposals for the Apollo Next Generation Sample Analysis Pro-gram (ANGSA), and 9 teams were selected to study a subset of the unopened and frozen samples [1]. The first sample opened as part of the ANGSA pro-gram was drive tube 73002. This was originally a ~30 cm long, 4 cm diameter drive tube collected on a land-slide deposit near Lara Crater at the Apollo 17 landing site. It was part of a ~60 cm long double drive tube collected, and the bottom half of the tube (73001) was sealed under vacuum on the Moon [2]. Prior to opening sample 73002, the sample was imaged with a high resolution X-ray Computed Tomography (XCT) scan of the entire tube [3], which provided invaluable information during the dissection process [4]. In addition to the pre-dissection XCT scans, individual >4 mm particles were separated from the 73002 regolith during processing and scanned at high resolution by XCT. Here we pre-sent the initial lithologic classification of 134 individual >4 mm rock fragments separated from the 73002 core during the dissection process. Methodology: Drive tube 73002 was manually dissected in 0.5 cm depth intervals in three passes (Fig. 1) [4,5]. Each interval from pass 1 and 2 was sieved to <1 mm and >1 mm size fractions, and >1 mm particles were further manually subdivided into 1-2, 2-4, 4-10, and >10 mm size fractions. Pass 3 was not sieved, but >10 mm clasts were separated manually. Each 4-10 mm and >10 mm fragment was individually weighed, triply bagged in Teflon, and scanned by XCT. There are 60 rock fragments in Pass 1, 64 rock fragments in Pass 2 (from the 4-10 mm and >10 mm size fractions), and 8 rock fragments from Pass 3 (>10 mm size fraction). Each individually bagged rock fragment was scanned using the 180 kV nano-focus transmission source on the Nikon XTH 320 XCT system at NASA Johnson Space Center [6]. Scanning conditions varied considerably for individual particles, in large part be-cause of the large variation in size (0.008-19.623 g). All scans fell within the following range of scan conditions: 2.8-20.6 um voxel size; 90-155 kV voltage; 18-39 uA current; 1891-3141 projections; and 902-2000 slices. Results and Discussion: The 132 rock fragments from sample 73002 fall into the following general cat-gories: agglutinates (n = 6); basalts (13); impact melts (5); impact melt breccias (IMB; 42); regolith breccias (62); and soil breccias (4); see Figure 2 for representative examples of each lithology. Within most of these broad lithologic groups are recognizable subgroups. For example, a significant portion of regolith breccia fragments contain some agglutinate-like glass (n = 9) or are dilithologic (7) because they contain a single large clast (~50% or more by volume). Subgroups can be based on similarities to previously identified lunar lithologies, such as high-Ti basalts (9) and VLT basalts (4; Fig. 3), or based on commonly seen features within the fragments, such as poikilitic ilmenite IMB (10), ilmenite-lath IMB (12), or vesicular IMB (11) “groups”. Particles in the same “group” are not necessarily intended to be genetically related, but rather identify particles that are similar and that follow-up studies can classify in more detail [7]. Conclusion: Identification of lithologies based on XCT is a powerful tool, but a more absolute classification will sometimes require additional textural information from thin sections (e.g., glassy-matrix regolith vs. impact-melt vs. granulitic breccia) or quantitative mineral compositions (e.g., basalt vs. monomict breccia).

R A Zeigler↗

An evaluation of the signature extension approach to large area crop inventories utilizing space image data

The author has identified the following significant results. Two examples of haze correction algorithms were tested: CROP-A and XSTAR. The CROP-A was tested in a unitemporal mode on data collected in 1973-74 over ten sample segments in Kansas. Because of the uniformly low level of haze present in these segments, no conclusion could be reached about CROP-A's ability to compensate for haze. It was noted, however, that in some cases CROP-A made serious errors which actually degraded classification performance. The haze correction algorithm XSTAR was tested in a multitemporal mode on 1975-76 LACIE sample segment data over 23 blind sites in Kansas and 18 sample segments in North Dakota, providing wide range of haze levels and other conditions for algorithm evaluation. It was found that this algorithm substantially improved signature extension classification accuracy when a sum-of-likelihoods classifier was used with an alien rejection threshold.

Nalepka, R. F.↗