Search NASA⌕ Search

SEARCH · Search NASA

Results for “data 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 901 records · Page 50

Coastal wetlands analysis from ERTS MSS digital data and field spectral measurements

Classification, delineation and evaluation of coastal wetlands can be made on the basis of major vegetative associations. To produce wetland maps, two vegetation-analysis look-up tables were developed for use in the ERTS ANALYSIS System. These look-up tables are based on Earth Resources Technology Satellite (ERTS) digital values in Multispectral Scanner (MSS) bands 4, 5, and 7 and were developed using seasonal spectral reflectance measurements from field observations. Computer-generated maps at an approximate scale of 1:20,000 were produced for the primary test site, Chincoteague Marsh, Virginia. There is a high degree of accuracy in the identification of wetland features and plant associations. The classification was also tested on other Atlantic coast salt marshes and a brackish marsh in the Chesapeake Bay.

Carter, V.↗

Single-class classification

Often, when classifying multispectral data, only one class or crop is of interest, such as wheat in the Large Area Crop Inventory Experiment (LACIE). Usual procedures for designing a Bayes classifier require that labeled training samples and therefore ground truth be available for the 'class of interest' plus all confusion classes defined by the multispectral data. This paper will consider the problem of designing a two-class Bayes classifier which will classify data into the 'class of interest' or the 'other' classes but will require only labeled training samples from the 'class of interest' to design the classifier. Thus, this classifier minimizes the need for ground truth. For these reasons, the classifier is referred to as a single-class classifier. A procedure for evaluating the overall performance of the single-class classifier in terms of the probability of error will be discussed.

Minter, T. C.↗

Texture transforms of remote sensing data

Tone and texture are fundamental interrelated visual concepts. The concepts are used for the digital analysis of remotely sensed image data. The reported investigation had the objective to develop software for the quantification of image texture and to apply the texture information to both image enhancement and thematic classification of remotely sensed data. The quantitative texture information was applied to the analysis of Landsat-2 Multispectral Scanner Subsystem (MSS) data. Attention is given to the characterization of image texture, textured transformations, the subtext program, and a description of methods and results. It is pointed out that the inability to use the texture transforms of the Landsat MSS data for the thematic mapping of the study area's land cover contrasts sharply with the reported results of the textural analysis of digitized aerial photography by Hsu (1978).

Irons, J. R.↗

Multivariate Density Estimation and Remote Sensing

Current efforts to develop methods and computer algorithms to effectively represent multivariate data commonly encountered in remote sensing applications are described. While this may involve scatter diagrams, multivariate representations of nonparametric probability density estimates are emphasized. The density function provides a useful graphical tool for looking at data and a useful theoretical tool for classification. This approach is called a thunderstorm data analysis.

Scott, D. W.↗

Analysis of Synthetic Aperture Radar data acquired over a variety of land cover

An analysis has been conducted of two-look-angle, multipolarization X-band SAR results. On the basis of the variety of land covers studied, the vertical-vertical polarization (VV) data is judged to contain the highest degree of contrast, while the horizontal-vertical (HV) polarization contained the least. VV polarization data is accordingly recommended for forest vegetation classification in those cases where only one data channel is available. The inclusion of horizontal-horizontal polarization data, however, is noted to be capable of delineating special surface features.

Wu, S. T.↗

Data supporting manuscript from L. Sheneman, G. Stephanopoulos, A.E. Vasdekis titled "Deep learning classification of lipid droplets in quantitative phase images" as currently under review at PLOS ONE. This includes: 1) raw and binary labeled Quantitative Phase Images (QPI) of Y. lipolytica cells used in the analyses described within the manuscript. 2) various derived data including classifier scores, etc.

Data supporting manuscript from L. Sheneman, G. Stephanopoulos, A.E. Vasdekis titled "Deep learning classification of lipid droplets in quantitative phase images" as currently under review at PLOS ONE. This includes: 1) raw and binary labeled Quantitative Phase Images (QPI) of Y. lipolytica cells used in the analyses described within the manuscript. 2) various derived data including classifier scores, etc.

ANN↗

The decision tree approach to classification

A class of multistage decision tree classifiers is proposed and studied relative to the classification of multispectral remotely sensed data. The decision tree classifiers are shown to have the potential for improving both the classification accuracy and the computation efficiency. Dimensionality in pattern recognition is discussed and two theorems on the lower bound of logic computation for multiclass classification are derived. The automatic or optimization approach is emphasized. Experimental results on real data are reported, which clearly demonstrate the usefulness of decision tree classifiers.

Wu, C.↗

Automated Stellar Spectra Classification with Ensemble Convolutional Neural Network

Large sky survey telescopes have produced a tremendous amount of astronomical data, including spectra. Machine learning methods must be employed to automatically process the spectral data obtained by these telescopes. Classification of stellar spectra by applying deep learning is an important research direction for the automatic classification of high-dimensional celestial spectra. In this paper, a robust ensemble convolutional neural network (ECNN) was designed and applied to improve the classification accuracy of massive stellar spectra from the Sloan digital sky survey. We designed six classifiers which consist six different convolutional neural networks (CNN), respectively, to recognize the spectra in DR16. Then, according the cross-entropy testing error of the spectra at different signal-to-noise ratios, we integrate the results of different classifiers in an ensemble learning way to improve the effect of classification. The experimental result proved that our one-dimensional ECNN strategy could achieve 95.0% accuracy in the classification task of the stellar spectra, a level of accuracy that exceeds that of the classical principal component analysis and support vector machine model.

79 ASTRONOMY AND ASTROPHYSICS↗

The Use of Thematic Mapper Data for Land Cover Discrimination: Preliminary Results from the UK Satmap Programme

In assessing the accuracy of classification techniques for Thematic Mapper data the consistency of the detector-to-detector response is critical. Preliminary studies were undertaken, therefore, to assess the significance of this factor for the TM. The overall structure of the band relationships can be examined by principal component analysis. In order to examine the utility of the Thematic Mapper data more carefully, six different land cover classes approximately Anderson level 1 were selected. These included an area of water from the sediment-laden Mississippi, woodland, agricultural land and urban land. A plume class was also selected which includes the plume of smoke emanating from the power station and drifting over the Mississippi river.

Jackson, M. J.↗

Improving ProtoDUNE pion cross-section measurements with NuGraph Michel-electron tagging

Understanding hadron-argon interactions is essential for precise neutrino energy reconstruction and final-state interaction modeling in liquid-argon time projection chamber (LArTPC) experiments such as DUNE. In particular, pion absorption and charge-exchange processes constitute significant sources of systematic uncertainty in neutrino oscillation measurements. ProtoDUNE-SP, a large-scale LArTPC prototype operated at the CERN Neutrino Platform and exposed to charged-particle test beams in the few-GeV range, enables direct measurements of these processes. This work focuses on the measurement of differential cross sections for pion absorption and charge exchange using the 2 GeV/c pion beam data from the ProtoDUNE-SP run. A key component of this analysis is the identification of Michel electrons from $\pi \rightarrow \mu \rightarrow e$ decay chains, which helps separate different interaction topologies and improves background rejection. Michel electron identification will also assist in reliably calibrating the electromagnetic response in ProtoDUNE-SP data and for the future DUNE detectors. In this analysis, we apply NuGraph to identify Michel electrons. NuGraph is a graph neural network that models detector hits as nodes connected by spatial and temporal edges for particle and topology classification in LArTPC detectors. We first benchmark NuGraph’s Michel electron classification performance using ICEBERG data, a small-scale LArTPC prototype used for DUNE electronics and reconstruction development, and then transfer the approach to ProtoDUNE-SP. This poster presents the analysis strategy, NuGraph-based classification studies, and discusses how these developments are expected to improve the pion cross-section measurement.

Razafinime, Soamasina Herilala [Cincinnati U.] (OR↗

Outage Cause Classification of Power Distribution Systems with Machine Learning and Real-World Data

Power distribution systems are geographically dispersed by nature. It may be affected by various factors, such as vegetation, weather, animal and human behaviors. Present response procedures to an outage event massively rely on expert experience and thus tend to be time-consuming. Automatic outage event detection and classification will help to reduce the responding and restoration time. However, this issue is less addressed with existing research done in this area. In this applied research, a set of waveform pre-processing techniques are first proposed to prepare the waveform data for being used as inputs to the classification algorithm. Further, a machine learning-based algorithm is proposed to classify the outage events according to their root causes, e.g. tree contact, animal contact, lightning, etc. Available data include three phase current & voltage waveforms and contextual information during the distribution system outages. The proposed machine learning algorithm takes the current and voltage waveforms as direct inputs in search of features that humans are unable to capture. Real data provided by a distribution company in the East Tennessee region is used to test the proposed pre-processing techniques and the classification algorithm.

Sun, Haoyuan↗

DETECTING VARIABILITY IN MASSIVE ASTRONOMICAL TIME-SERIES DATA. II. VARIABLE CANDIDATES IN THE NORTHERN SKY VARIABILITY SURVEY

We present variability analysis of data from the Northern Sky Variability Survey (NSVS). Using the clustering method, which defines variable candidates as outliers from large clusters, we cluster 16,189,040 light curves having data points at more than 15 epochs as variable and non-variable candidates in 638 NSVS fields. Variable candidates are selected depending on how strongly they are separated from the largest cluster and how rarely they are grouped together in eight-dimensional space spanned by variability indices. All NSVS light curves are also cross-correlated with IRAS, AKARI, Two Micron All Sky Survey, Sloan Digital Sky Survey (SDSS), and GALEX objects, as well as known objects in the SIMBAD database. The variability analysis and cross-correlation results are provided in a public online database, which can be used to select interesting objects for further investigation. Adopting conservative selection criteria for variable candidates, we find about 1.8 million light curves as possible variable candidates in the NSVS data, corresponding to about 10% of our entire NSVS sample. Multi-wavelength colors help us find specific types of variability among the variable candidates. Moreover, we also use morphological classification from other surveys such as SDSS to suppress spurious cases caused by blending objects or extended sources due to the low angular resolution of the NSVS.

79 ASTRONOMY AND ASTROPHYSICS↗

Potentially efficient forest and range applications of remote sensing using earth orbital space craft, circa 1980

Sixteen remote sensing applications or groups of related applications judged to be most important of any in the forestry and range disciplines were evaluated. In one application, major land classification, large amounts of useful data are anticipated to be contributed by space sensors in 1980. In four applications moderate amounts are anticipated to be so contributed. These are timber inventory, range inventory, fire weather forecasting, and monitoring snowfields. In the following seven applications small but significant amounts of data are anticipated to be contributed by space sensors: (1) detailed land classification; (2) inventory of wildlife habitat; (3) recreation resource inventory; (4) detecting stresses on the vegetation (5) monitoring air pollution caused by wildfires and prescribed burning; (6) monitoring water cycle, (7) pollution and erosion; and (8) evaluating damage to forests and ranges.

Wilson, R. C.↗

Polarization effects and multipolarization SAR

Imaging radar polarimeters are usually implemented using a Synthetic Aperture Radar (SAR) approach to give a high resolution image in two dimensions: range and azimuth. For each pixel in the image a polarimetric SAR gives sufficient information to characterize the polarimetric scattering properties of the imaged area (or target) as seen by the radar. Using a polarimetric SAR system as opposed to a single-polarization SAR system provides significantly more information about the target scattering mechanisms and allows better discrimination between different types of surfaces. In these notes a brief overview of SAR polarimetry is offered. The notes are intended as a text to accompany a lecture on SAR polarimetry as part of an AGARD-NATO course. Covered in the notes are the following: the polarization properties of electromagnetic waves; the concepts of radar scattering and measuring radar backscatter with a SAR; polarization synthesis; scattering matrix, Stokes matrix, and covariance matrix representations of polarimetric SAR data; polarization signature plots; design and calibration of polarimetric SAR systems; polarization filtering for target detection; fitting a simple model to polarimetric SAR measurements of naturally occurring features; and supervised classification of polarimetric SAR data.

Freeman, Anthony↗

Modeling Spatial Dependencies and Semantic Concepts in Data Mining

Data mining is the process of discovering new patterns and relationships in large datasets. However, several studies have shown that general data mining techniques often fail to extract meaningful patterns and relationships from the spatial data owing to the violation of fundamental geospatial principles. In this tutorial, we introduce basic principles behind explicit modeling of spatial and semantic concepts in data mining. In particular, we focus on modeling these concepts in the widely used classification, clustering, and prediction algorithms. Classification is the process of learning a structure or model (from user given inputs) and applying the known model to the new data. Clustering is the process of discovering groups and structures in the data that are ``similar,'' without applying any known structures in the data. Prediction is the process of finding a function that models (explains) the data with least error. One common assumption among all these methods is that the data is independent and identically distributed. Such assumptions do not hold well in spatial data, where spatial dependency and spatial heterogeneity are a norm. In addition, spatial semantics are often ignored by the data mining algorithms. In this tutorial we cover recent advances in explicitly modeling of spatial dependencies and semantic concepts in data mining.

99 GENERAL AND MISCELLANEOUS↗

Comparative evaluation of spatial features in automatic land use classification from photographic imagery

Five spatial features have been evaluated for their applicability in automatic land use classification of photographic imagery. Data arrays of approximately 10,000 square meters were classified on the basis of an 8 by 8 point data array. Statistical features, information features, sequency features and texture features were evaluated using a distance to prototype classifier and an adaptive classifier. The results indicate approximately 70% accuracy in the classification.

Herzog, J. H.↗

Development of a data management front end for use with a LANDSAT based information system

A user friendly front end was constructed to facilitate access to the LANDSAT mosaic data base supplied by JPL and to process both LANDSAT and ancillary data. Archieval and retrieval techniques were developed to efficiently handle this data base and make it compatible with requirements of the Pennsylvania Bureau of Forestry. Procedures are ready for: (1) forming the forest/nonforest mask in ORSER compressed map format using GSFC-supplied classification procedures; (2) registering data from a new scene (defoliated) to the mask (which may involve mosaicking if the area encompasses two LANDSAT scenes; (3) producing a masked new data set using the MASK program; (4) analyzing this data set to produce a map showing degrees of defoliation, output on the Versatec plotter; and (5) producing color composite maps by a diazo-type process.

Turner, B. J.↗

Classification and area estimation of land covers in Kansas using ground-gathered and LANDSAT digital data

Ground-gathered data and LANDSAT multispectral scanner (MSS) digital data from 1981 were analyzed to produce a classification of Kansas land areas into specific types called land covers. The land covers included rangeland, forest, residential, commercial/industrial, and various types of water. The analysis produced two outputs: acreage estimates with measures of precision, and map-type or photo products of the classification which can be overlaid on maps at specific scales. State-level acreage estimates were obtained and substate-level land cover classification overlays and estimates were generated for selected geographical areas. These products were found to be of potential use in managing land and water resources.

May, G. A.↗