Results for “Image Classification” — Page 63

NASA NTRSDate not supplied

Mapping Inundation from Hurricane Florence (2018) with L-Band Synthetic Aperture Radar, Commercial Imagery, and Ancillary Data via Random Forest Classification

Mapping the extent of floodwaters following extreme rainfall aids in the distribution of resources, recovery efforts, and damage assessment practices. Development of a land cover classification system focused on mapping inundation after major hurricane events using synthetic aperture radar (SAR) data could allow for the production of near-real-time inundation mapping, enabling government and emergency response entities to get a preliminary idea of a developing situation. In response to Hurricane Florence of 2018, NASA JPL collected numerous swaths of quad-pol L-band SAR data with the Uninhabited Aerial Vehicle Synthetic Aperture Radar (UAVSAR) instrument observing the record-setting river stages across North and South Carolina. The resulting fully-polarized SAR images allow for mapping of inundation extent at a high spatial resolution with a unique advantage over optical imaging stemming from the sensor’s ability to penetrate cloud cover and dense vegetation. This study seeks to determine how accurately maps of inundation can be generated from L-band SAR imagery through Random Forest classification. Once the extent of water and inundated vegetation is classified, cleanup operations are performed using fuzzy logic to reduce false detections. Estimates of water extent are then combined with datasets describing the distribution of population, buildings, and roads throughout the domain to evaluate societal impacts. Results from the Hurricane Florence case study will be discussed along with the limitations of available validation data for assessment of the classifier’s accuracy.

Alexander Melancon↗

DOE OSTI2026

A machine learning pipeline for identifying infiltration managed aquifer recharge locations from satellite imagery in the San Joaquin Valley, California

This study focuses on an agricultural region in California’s Central Valley, USA, where Managed Aquifer Recharge (MAR) is widely implemented to mitigate groundwater depletion under increasing water demand and climate variability. A deep learning and machine learning framework was developed to identify infiltration-MAR locations using satellite imagery and environmental data. The framework integrates surface water detection from Sentinel-2 imagery, geospatial delineation of water bodies, spatiotemporal tracking of water body dynamics, and supervised classification using meteorological, environmental, and topographic variables. The framework was applied to a 2379 km² study area southwest of Fresno, where 765 water bodies were detected, including 139 identified MAR sites based on publicly available datasets and expert knowledge. The classification model achieved an accuracy of 0.94 and an F1 score of 0.85. Feature importance analysis indicates that cropland, normalized difference vegetation index (NDVI), and evaporation are among the most influential predictors for infiltration-MAR. Notably, the framework suggests that engineered water management in infiltration-MAR systems can disrupt or even reverse the expected positive correlation between surface water extent and precipitation. These findings provide physically interpretable insights into the characteristics of existing infiltration-MAR facilities and demonstrate the potential of the proposed framework as a reproducible, interpretable, and potentially transferable tool for data-driven infiltration-MAR identification and inventory development under growing climatic and hydrological uncertainty.

Classification↗

NASA NTRS1983

Remote sensing of agricultural crops and soils

Research in the correlative and noncorrelative approaches to image registration and the spectral estimation of corn canopy phytomass and water content is reported. Scene radiation research results discussed include: corn and soybean LANDSAT MSS classification performance as a function of scene characteristics; estimating crop development stages from MSS data; the interception of photosynthetically active radiation in corn and soybean canopies; costs of measuring leaf area index of corn; LANDSAT spectral inputs to crop models including the use of the greenness index to assess crop stress and the evaluation of MSS data for estimating corn and soybean development stages; field research experiment design data acquisition and preprocessing; and Sun-view angles studies of corn and soybean canopies in support of vegetation canopy reflection modeling.

Bauer, M. E.↗

NASA NTRS1990

Processing Of Multispectral Data For Identification Of Rocks

Linear discriminant analysis and supervised classification evaluated. Report discusses processing of multispectral remote-sensing imagery to identify kinds of sedimentary rocks by spectral signatures in geological and geographical contexts. Raw image data are spectra of picture elements in images of seven sedimentary rock units exposed on margin of Wind River Basin in Wyoming. Data acquired by Landsat Thematic Mapper (TM), Thermal Infrared Multispectral Scanner (TIMS), and NASA/JPL airborne synthetic-aperture radar (SAR).

Evans, Diane L.↗

NASA NTRS1990

Multiband radar characterization of forest biomes

The utility of airborne and orbital SAR in classification, assessment, and monitoring of forest biomes is investigated through analysis of orbital synthetic aperature radar (SAR) and multifrequency and multipolarized airborne SAR imagery relying on image tone and texture. Preliminary airborne SAR experiments and truck-mounted scatterometer observations demonstrated that the three dimensional structural complexity of a forest, and the various scales of temporal dynamics in the microwave dielectric properties of both trees and the underlying substrate would severely limit empirical or semi-empirical approaches. As a consequence, it became necessary to develop a more profound understanding of the electromagnetic properties of a forest scene and their temporal dynamics through controlled experimentation coupled with theoretical development and verification. The concatenation of various models into a physically-based composite model treating the entire forest scene became the major objective of the study as this is the key to development of a series of robust retrieval algorithms for forest biophysical properties. In order to verify the performance of the component elements of the composite model, a series of controlled laboratory and field experiments were undertaken to: (1) develop techniques to measure the microwave dielectric properties of vegetation; (2) relate the microwave dielectric properties of vegetation to more readily measured characteristics such as density and moisture content; (3) calculate the radar cross-section of leaves, and cylinders; (4) improve backscatter models for rough surfaces; and (5) relate attenuation and phase delays during propagation through canopies to canopy properties. These modeling efforts, as validated by the measurements, were incorporated within a larger model known as the Michigan Microwave Canopy Scattering (MIMICS) Model.

Dobson, M. Craig↗

NASA NTRS1998

Characteristics of Forests in Western Sayani Mountains, Siberia from SAR Data

This paper investigated the possibility of using spaceborne radar data to map forest types and logging in the mountainous Western Sayani area in Siberia. L and C band HH, HV, and VV polarized images from the Shuttle Imaging Radar-C instrument were used in the study. Techniques to reduce topographic effects in the radar images were investigated. These included radiometric correction using illumination angle inferred from a digital elevation model, and reducing apparent effects of topography through band ratios. Forest classification was performed after terrain correction utilizing typical supervised techniques and principal component analyses. An ancillary data set of local elevations was also used to improve the forest classification. Map accuracy for each technique was estimated for training sites based on Russian forestry maps, satellite imagery and field measurements. The results indicate that it is necessary to correct for topography when attempting to classify forests in mountainous terrain. Radiometric correction based on a DEM (Digital Elevation Model) improved classification results but required reducing the SAR (Synthetic Aperture Radar) resolution to match the DEM. Using ratios of SAR channels that include cross-polarization improved classification and

Ranson, K. Jon↗

NASA NTRS2020

Comparison of Cloud Detection Algorithms for Sentinel-2 Imagery

Accurate, automated cloud and cloud shadow detection is a key component of the processing needed to prepare optical satellite imagery for scientific analysis. Many existing cloud detection algorithms rely on temperature information to identify clouds, making detection difficult for imagers that lack a thermal band, like Sentinel-2. To get maximum benefit from Sentinel-2 products it is critical to understand which algorithms best identify clouds and their shadows in images. We examined the relative performance of five different cloud-masking algorithms (Sen2Cor, MAJA, LaSRC, Fmask and Tmask) in 6 Sentinel-2 scenes (28 total images) distributed across the Eastern Hemisphere. Expanding on these comparisons, we tested ensemble approaches to improve results. We tested three ensemble approaches to cloud and shadow classification based on the outputs of the five initial algorithms using the cloud masks in: (1) a majority prediction model; (2) a random forests model; and (3) a conditional logic model. Accuracy assessments show a trade-off between omission and commission errors in cloud detection for individual algorithms across all sites, and some algorithms are better at detecting either clouds or cloud shadows. No single algorithm outperforms the others for both clouds and shadows. Aggregating the results from multiple algorithms produces fewer undetected clouds and higher overall accuracy than any single algorithm, with as high as 2.7% improvement over the top-performing algorithm, suggesting an ensemble approach may be the most useful for processing of Sentinel-2 data.

Sentinel-2↗

NASA NTRSDate not supplied

Mapping Aerosol Lidar Ratios Over Ocean for CALIPSO Using Modis AOD-Constrained Retrievals and A Global Aerosol Model

The CALIPSO aerosol algorithms currently assign one lidar ratio value for each of the seven tropospheric aerosol types. In a future data products release, the CALIPSO project aims to improve these algorithms by developing regional and seasonal lidar ratio climatologies. In this study, aerosol lidar ratios are inferred through Cloud-Aerosol Lidar with Orthogonal Polarization (CALIOP) backscatter profiles constrained by collocated aerosol optical depth from Aqua Moderate Resolution Imaging Spectroradiometer (MODIS) datasets. The analysis is subsampled for only those profiles that are cloud-free and contain one aerosol type (based on CALIOP feature classification). In addition, the CALIOP profiles are collocated with aerosol volume fractions simulated by the Goddard Chemistry Aerosol Radiation and Transport (GOCART) model. In this talk, the twelve-year (2006-2017) mean spatial distributions of inferred aerosol lidar ratios for CALIOP-classified marine aerosols and the corresponding modeled sea salt volume fractions (SSVF) will be shown. Model-assisted climatological lidar ratio maps on a seasonal scale will also be presented, developed from the empirical relationship found between the modeled SSVF and lidar ratios. A comparison with past studies will be provided as well as results of a sensitivity study regarding the variability of retrieved lidar ratios as a function of horizontal averaging resolution. While the majority of this talk will focus on lidar ratios for CALIOP-classified marine aerosols, preliminary results will be shown for other aerosol types over ocean, such as dust and elevated smoke.

Travis Toth↗

NASA NTRS2013

Classification of Dust Days by Satellite Remotely Sensed Aerosol Products

Considerable progress in satellite remote sensing (SRS) of dust particles has been seen in the last decade. From an environmental health perspective, such an event detection, after linking it to ground particulate matter (PM) concentrations, can proxy acute exposure to respirable particles of certain properties (i.e. size, composition, and toxicity). Being affected considerably by atmospheric dust, previous studies in the Eastern Mediterranean, and in Israel in particular, have focused on mechanistic and synoptic prediction, classification, and characterization of dust events. In particular, a scheme for identifying dust days (DD) in Israel based on ground PM10 (particulate matter of size smaller than 10 nm) measurements has been suggested, which has been validated by compositional analysis. This scheme requires information regarding ground PM10 levels, which is naturally limited in places with sparse ground-monitoring coverage. In such cases, SRS may be an efficient and cost-effective alternative to ground measurements. This work demonstrates a new model for identifying DD and non-DD (NDD) over Israel based on an integration of aerosol products from different satellite platforms (Moderate Resolution Imaging Spectroradiometer (MODIS) and Ozone Monitoring Instrument (OMI)). Analysis of ground-monitoring data from 2007 to 2008 in southern Israel revealed 67 DD, with more than 88 percent occurring during winter and spring. A Classification and Regression Tree (CART) model that was applied to a database containing ground monitoring (the dependent variable) and SRS aerosol product (the independent variables) records revealed an optimal set of binary variables for the identification of DD. These variables are combinations of the following primary variables: the calendar month, ground-level relative humidity (RH), the aerosol optical depth (AOD) from MODIS, and the aerosol absorbing index (AAI) from OMI. A logistic regression that uses these variables, coded as binary variables, demonstrated 93.2 percent correct classifications of DD and NDD. Evaluation of the combined CART-logistic regression scheme in an adjacent geographical region (Gush Dan) demonstrated good results. Using SRS aerosol products for DD and NDD, identification may enable us to distinguish between health, ecological, and environmental effects that result from exposure to these distinct particle populations.

satellite remote sensing↗

NASA NTRS2017

Linear Subpixel Learning Algorithm for Land Cover Classification from WELD using High Performance Computing

In this work, we use a Fully Constrained Least Squares Subpixel Learning Algorithm to unmix global WELD (Web Enabled Landsat Data) to obtain fractions or abundances of substrate (S), vegetation (V) and dark objects (D) classes. Because of the sheer nature of data and compute needs, we leveraged the NASA Earth Exchange (NEX) high performance computing architecture to optimize and scale our algorithm for large-scale processing. Subsequently, the S-V-D abundance maps were characterized into 4 classes namely, forest, farmland, water and urban areas (with NPP-VIIRS-national polar orbiting partnership visible infrared imaging radiometer suite nighttime lights data) over California, USA using Random Forest classifier. Validation of these land cover maps with NLCD (National Land Cover Database) 2011 products and NAFD (North American Forest Dynamics) static forest cover maps showed that an overall classification accuracy of over 91 percent was achieved, which is a 6 percent improvement in unmixing based classification relative to per-pixel-based classification. As such, abundance maps continue to offer an useful alternative to high-spatial resolution data derived classification maps for forest inventory analysis, multi-class mapping for eco-climatic models and applications, fast multi-temporal trend analysis and for societal and policy-relevant applications needed at the watershed scale.

Subpixel↗

NASA NTRS1984

Assessment of computer techniques for processing digital LANDSAT MSS data for lithological discrimination of Serra do Ramalho, State of Bahia

Enhancement techniques and thematic classifications were applied to the metasediments of Bambui Super Group (Upper Proterozoic) in the Region of Serra do Ramalho, SW of the state of Bahia. Linear contrast stretch, band-ratios with contrast stretch, and color-composites allow lithological discriminations. The effects of human activities and of vegetation cover mask and limit, in several ways, the lithological discrimination with digital MSS data. Principal component images and color composite of linear contrast stretch of these products, show lithological discrimination through tonal gradations. This set of products allows the delineations of several metasedimentary sequences to a level superior to reconnaissance mapping. Supervised (maximum likelihood classifier) and nonsupervised (K-Means classifier) classification of the limestone sequence, host to fluorite mineralization show satisfactory results.

Paradella, W. R.↗

NASA NTRS2020

Automated Cardiovascular Pathology Assessment using Semantic Segmentation and Ensemble Learning

Cardiac magnetic resonance imaging provides high spatial resolution, enabling improved extraction of important functional and morphological features for cardiovascular disease staging. Segmentation of ventricular cavities and myocardium in cardiac cine sequencing provides a basis to quantify cardiac measures such as ejection fraction. A method is presented that curtails the expense and observer bias of manual cardiac evaluation by combining semantic segmentation and disease classification into a fully automatic processing pipeline. The initial processing element consists of a robust dilated convolutional neural network architecture for voxel-wise segmentation of the myocardium and ventricular cavities. The resulting comprehensive volumetric feature matrix captures diagnostic clinical procedure data and is utilized by the final processing element to model a cardiac pathology classifier. Our approach evaluated anonymized cardiac images from a training data set of 100 patients (4 pathology groups, 1 healthy group, 20 patients per group) examined at the University Hospital of Dijon. The top average Dice index scores achieved were 0.940, 0.886, 0.849 for structure segmentation of the left ventricle (LV), myocardium and right ventricle (RV) respectively. A 5-ary pathology classification accuracy of 90% was recorded on an independent test set using the trained model. Performance results demonstrate potential for advanced machine learning methods to deliver accurate, efficient and reproducible cardiac pathological assessment.

Semantic Segmentation↗

NASA NTRS2019

Flood Mapping of Recent Major Hurricane Events with Synthetic Aperture Radar, Commercial Imaging, and Aerial Observations

Floodwater mapping is an important remote sensing process that is used for disaster response, recovery, and damage assessment practices. Developing a system to read in Synthetic Aperture Radar (SAR) data and perform land cover classification will allow for the production of near real-time inundation mapping, enabling government and emergency response entities to get a preliminary idea of the situation. SAR is a unique remote sensing tool. Data in this project was obtained by NASA Jet Propulsion Laboratory’s Uninhabited Aerial Vehicle SAR (UAVSAR), an L-band radar mounted to a Gulfstream III jet. Data collected by UAVSAR is similar to what will be available from the NASA-Indian Space Research Organization (NISAR) mission starting in early 2022. Using Python and ArcGIS applications, a model was developed using training samples taken from NOAA post-event aerial photography and UAVSAR data gathered in the aftermath of Hurricane Florence in September 2018.

Melancon, Alexander M.↗

NASA NTRSDate not supplied

Using Machine Learning to Infer Material Properties of Debris Fragments from X-ray Images in the DebriSat Project

The DebriSat project is a collaboration effort with the NASA Orbital Debris Program Office, the U.S. Space Force Space Systems Command Center, The Aerospace Corporation, and the University of Florida. To date, over 200,000 fragments from this ground-based, hypervelocity impact experiment have been collected, and processing is underway to determine their physical characteristics, such as material, shape, color, characteristic length, and average cross-sectional area. The x-ray process is primarily used to identify the location of the fragments and estimated size for extraction, so that these physical characteristics can be assessed. This paper proposes a machine learning-based approach to characterize materials from x-ray images of debris fragments embedded in soft-catch foam used in the DebriSat project. The novel methodology discussed in this paper will highlight the use of x-ray imagery data to characterize these fragments without extraction or a human-in-the-loop. Both supervised and unsupervised machine learning techniques are utilized with this approach to infer the physical parameters of the fragments embedded in the soft-catch foam panels used in the impact experiment based on x-ray images of the foam panels. Additionally, 3D reconstructions of the extracted fragments are created with images taken from two different angles using the structure from motion (SfM) method. The characteristic lengths and shape from the 3D reconstruction, alongside the physical characteristics of the debris, are used in the inference of the material type. To develop and test the approach, a dataset of x-ray images of debris fragments of varying sizes and materials is collected. Supervised learning methods such as convolutional neural networks (CNNs), support vector machines (SVM), decision trees, and random forest classifiers are used due to the high-dimensional feature spaces of the debris and nonlinear decision boundaries for material categorization. Given the limited pre-labeled data of embedded debris materials smaller than 10 mm, unsupervised machine learning techniques such as clustering algorithms and autoencoders are used, in addition to supervised learning methods. The clustering algorithms group similar fragments together based on their physical properties, and autoencoders reduce the dimensionality of the x ray images and extract relevant features. The performance of the proposed approach's is analyzed using a range of statistical methods, including confusion matrices, receiver operating characteristic curves, and precision-recall curves. The results are compared with those obtained using a baseline approach that relies on manual identification and classification of debris fragments. To evaluate the effectiveness of different machine learning methods, statistical tests such as t-tests, ANOVA, and cross-validation are performed, comparing the performance of CNNs, SVMs, clustering algorithms, and autoencoders. Additional analysis needs to be conducted to identify any sources of bias or variability that may affect the results, such as variations in imaging conditions or fragmentation patterns. Other topics explored are limitations, refinements, and the potential use of semi-supervised learning techniques, such as self-training to label unlabeled datasets and co-training using x-ray images taken from two different angles as two different models.

Saik Anam Siam↗

NASA NTRS2023

Using Machine Learning to Infer Material Properties of Debris Fragments from X-ray Images in the DebriSat Project

The DebriSat project is a collaboration effort with the NASA Orbital Debris Program Office, the U.S. Space Force Space Systems Command Center, The Aerospace Corporation, and the University of Florida. To date, over 200,000 fragments from this ground-based, hypervelocity impact experiment have been collected, and processing is underway to determine their physical characteristics, such as material, shape, color, characteristic length, and average cross-sectional area. The x-ray process is primarily used to identify the location of the fragments and estimated size for extraction, so that these physical characteristics can be assessed. This paper proposes a machine learning-based approach to characterize materials from x-ray images of debris fragments embedded in soft-catch foam used in the DebriSat project. The novel methodology discussed in this paper will highlight the use of x-ray imagery data to characterize these fragments without extraction or a human-in-the-loop. Both supervised and unsupervised machine learning techniques are utilized with this approach to infer the physical parameters of the fragments embedded in the soft-catch foam panels used in the impact experiment based on x-ray images of the foam panels. Additionally, 3D reconstructions of the extracted fragments are created with images taken from two different angles using the structure from motion (SfM) method. The characteristic lengths and shape from the 3D reconstruction, alongside the physical characteristics of the debris, are used in the inference of the material type. To develop and test the approach, a dataset of x-ray images of debris fragments of varying sizes and materials is collected. Supervised learning methods such as convolutional neural networks (CNNs), support vector machines (SVM), decision trees, and random forest classifiers are used due to the high-dimensional feature spaces of the debris and nonlinear decision boundaries for material categorization. Given the limited pre-labeled data of embedded debris materials smaller than 10 mm, unsupervised machine learning techniques such as clustering algorithms and autoencoders are used, in addition to supervised learning methods. The clustering algorithms group similar fragments together based on their physical properties, and autoencoders reduce the dimensionality of the x ray images and extract relevant features. The performance of the proposed approach's is analyzed using a range of statistical methods, including confusion matrices, receiver operating characteristic curves, and precision-recall curves. The results are compared with those obtained using a baseline approach that relies on manual identification and classification of debris fragments. To evaluate the effectiveness of different machine learning methods, statistical tests such as t-tests, ANOVA, and cross-validation are performed, comparing the performance of CNNs, SVMs, clustering algorithms, and autoencoders. Additional analysis needs to be conducted to identify any sources of bias or variability that may affect the results, such as variations in imaging conditions or fragmentation patterns. Other topics explored are limitations, refinements, and the potential use of semi-supervised learning techniques, such as self-training to label unlabeled datasets and co-training using x-ray images taken from two different angles as two different models.

Saik Anam Siam↗

DOE OSTI2024

Digital twins and deep learning segmentation of defects in monolayer MX 2 phases

Developing methods to understand and control defect formation in nanomaterials offers a promising route for materials discovery. Monolayer MX 2 phases represent a particularly compelling case for defect engineering of nanomaterials due to the large variability in their physical properties as different defects are introduced into their structure. However, effective identification and quantification of defects remain a challenge even as high-throughput scanning transmission electron microscopy methods improve. This study highlights the benefits of employing first principles calculations to produce digital twins for training deep learning segmentation models for defect identification in monolayer MX 2 phases. Around 600 defect structures were obtained using density functional theory calculations, with each monolayer MX 2 structure being subjected to multislice simulations for the purpose of generating the digital twins. Several deep learning segmentation architectures were trained on this dataset, and their performances evaluated under a variety of conditions such as recognizing defects in the presence of unidentified impurities, beam damage, grain boundaries, and with reduced image quality from low electron doses. Further, this digital twin approach allows benchmarking different deep learning architectures on a theory dataset, which enables the study of defect classification under a broad array of finely controlled conditions. It thus opens the door to resolving the underpinning physical reasons for model shortcomings and potentially chart paths forward for automated discovery of materials defect phases in experiments.

36 MATERIALS SCIENCE↗

NASA NTRS1974

Combined spectral and spatial processing of ERTS imagery data

A general procedure is presented for extracting textural properties of blocks of image data. These features are calculated in the spatial domain, taking into account the statistical nature of texture. The procedure is based on the assumption that the texture information in an image is contained in the overall or 'average' spatial relationship which the gray tones in the image have to one another. The spatial gray-tone dependence matrix is discussed together with the spectral features used in the study and the land use classification studies conducted. The results of the study show the usefulness of using both spectral and textural characteristics of ERTS multispectral scanner data for developing classification procedures.

Haralick, R. M.↗

NASA NTRS1974

The ERTS-1 investigation (ER-600). Volume 2: ERTS-1 coastal/estuarine analysis

The Coastal Analysis Team of the Johnson Space Center conducted a 1-year investigation of ERTS-1 MSS data to determine its usefulness in coastal zone management. Galveston Bay, Texas, was the study area for evaluating both conventional image interpretation and computer-aided techniques. There was limited success in detecting, identifying and measuring areal extent of water bodies, turbidity zones, phytoplankton blooms, salt marshes, grasslands, swamps, and low wetlands using image interpretation techniques. Computer-aided techniques were generally successful in identifying these features. Aerial measurement of salt marshes accuracies ranged from 89 to 99 percent. Overall classification accuracy of all study sites was 89 percent for Level 1 and 75 percent for Level 2.

Erb, R. B.↗