A data-driven approach for scour detection around monopile-supported offshore wind turbines using Naive Bayes classification
Not Available
Jawalageri, Satish (ORCID:0000000297829661)↗
SEARCH · Engineering Papers
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.
Continual Learning for Pattern Recognizers using Neurogenesis Deep Learning
Deep neural networks have emerged as a leading set of algorithms to infer information from a variety of data sources such as images and time series data. In their most basic form, neural networks lack the ability to adapt to new classes of information. Continual learning is a field of study attempting to give previously trained deep learning models the ability to adapt to a changing environment. Previous work developed a CL method called Neurogenesis for Deep Learning (NDL). Here, we combine NDL with a specific neural network architecture (the Ladder Network) to produce a system capable of automatically adapting a classification neural network to new classes of data. The NDL Ladder Network was evaluated against other leading CL methods. While the NDL and Ladder Network system did not match the cutting edge performance achieved by other CL methods, in most cases it performed comparably and is the only system evaluated that can learn new classes of information with no human intervention.
97 MATHEMATICS AND COMPUTING↗
Crop identification and acreage measurement utilizing ERTS imagery
There are no author-identified significant results in this report. The microdensitometer will be used to analyze data acquired by ERTS-1 imagery. The classification programs and software packages have been acquired and are being prepared for use with the information as it is received. Photo and digital tapes have been acquired for coverage of virtually 100 percent of the test site areas. These areas are located in South Dakota, Idaho, Missouri, and Kansas. Hass 70mm color infrared, infrared, black and white high altitude aerial photography of the test sites is available. Collection of ground truth for updating the data base has been completed and a computer program written to count the number of fields and give total acres by size group for the segments in each test site. Results are given of data analysis performed on digitized data from densitometer measurements of fields of corn, sugar, beets, and alfalfa in Kansas.
Vonsteen, D. H.↗
Long-Term Impacts of Selective Logging on Amazon Forest Dynamics from Multi-Temporal Airborne LiDAR
Forest degradation is common in tropical landscapes, but estimates of the extent and duration of degradation impacts are highly uncertain. In particular, selective logging is a form of forest degradation that alters canopy structure and function, with persistent ecological impacts following forest harvest. In this study, we employed airborne laser scanning in 2012 and 2014 to estimate three-dimensional changes in the forest canopy and understory structure and aboveground biomass following reduced-impact selective logging in a site in Eastern Amazon. Also, we developed a binary classification model to distinguish intact versus logged forests. We found that canopy gap frequency was significantly higher in logged versus intact forests even after 8 years (the time span of our study). In contrast, the understory of logged areas could not be distinguished from the understory of intact forests after 6–7 years of logging activities. Measuring new gap formation between LiDAR acquisitions in 2012 and 2014, we showed rates 2 to 7 times higher in logged areas compared to intact forests. New gaps were spatially clumped with 76 to 89% of new gaps within 5 m of prior logging damage. The biomass dynamics in areas logged between the two LiDAR acquisitions was clearly detected with an average estimated loss of -4.14 +/- 0.76 MgC/hay. In areas recovering from logging prior to the first acquisition, we estimated biomass gains close to zero. Together, our findings unravel the magnitude and duration of delayed impacts of selective logging in forest structural attributes, confirm the high potential of airborne LiDAR multitemporal data to characterize forest degradation in the tropics, and present a novel approach to forest classification using LiDAR data.
Pinage, Ekena Rangel↗
Venus: Global distribution of Pioneer-Venus radar roughness and reflectivity and global correlation of altimetry, roughness and reflectivity observations
An improved version of the Pionner-Venus orbital data was used for a statistical analysis of global radar roughness and (alpha(0)) (rho) reflectivity. Classification maps of the venusian surface are produced in a supervised manner on the basis of statistical and empirical studies of the individual data sets. The primary objective is to assess the degree of homogeneity of surface radar properties within topographic provinces in order to map possible geologic boundaries. Maps were produced by correlating two data sets at a time. Classification of specific regions, such as Ishtar, has demonstrated that distinct geological units can be identified.
Head, J. W.↗
Assessing Hurricane Katrina Vegetation Damage at Stennis Space Center using IKONOS Image Classification Techniques
Hurricane Katrina inflicted widespread damage to vegetation in southwestern coastal Mississippi upon landfall on August 29, 2005. Storm damage to surface vegetation types at the NASA John C. Stennis Space Center (SSC) was mapped and quantified using IKONOS data originally acquired on September 2, 2005, and later obtained via a Department of Defense ClearView contract. NASA SSC management required an assessment of the hurricane s impact to the 125,000-acre buffer zone used to mitigate rocket engine testing noise and vibration impacts and to manage forestry and fire risk. This study employed ERDAS IMAGINE software to apply traditional classification techniques to the IKONOS data. Spectral signatures were collected from multiple ISODATA classifications of subset areas across the entire region and then appended to a master file representative of major targeted cover type conditions. The master file was subsequently used with the IKONOS data and with a maximum likelihood algorithm to produce a supervised classification later refined using GIS-based editing. The final results enabled mapped, quantitative areal estimates of hurricane-induced damage according to general surface cover type. The IKONOS classification accuracy was assessed using higher resolution aerial imagery and field survey data. In-situ data and GIS analysis indicate that the results compare well to FEMA maps of flooding extent. The IKONOS classification also mapped open areas with woody storm debris. The detection of such storm damage categories is potentially useful for government officials responsible for hurricane disaster mitigation.
Spruce, Joseph P.↗
LAMMR world data base documentation support and demonstrations
The primary purpose of the World Surface Map is to provide the LAMMR subsystem with world surface type classifications that are used to set up LAMMR LEVEL II process control. This data base will be accessed solely by the LAMMR subsystem. The SCATT and ALT subsystems will access the data base indirectly through the T sub b (Brightness Temperature) Data Bank, where the surface types were updated from a priori to current classification, and where the surface types were organized on an orbital subtrack basis. The single most important factor in the design of the World Surface Maps is the ease of access to the information while the complexity of generating these maps is of lesser importance because their generation is a one-time, off-line process. The World Surface Map provides storage of information with a resolution of 7 km necessary to set flags concerning the earth's features with a different set of maps for each month of the year.
Chin, R.↗
Studies of optical and biological properties of terrestrial land cover using multispectral linear array technology
A series of experiments to study the optical and biological properties of terrestrial land cover are planned for late 1987 using a six-channel imaging spectroradiometer based on newly developed multispectral linear array (MLA) detector technology. Data from selected portions of the Sahel and rain forests of Africa and South America will be used to delineate biomass classes and estimate spherical albedos. A spatial resolution of 15 meters in the four visible-near IR channels and 30 meters in two shortwave IR channels, including a 'new' channel centered at 1.24 micrometers when combined with a spectral width of 20 nm for all channels, will be used to investigate possible improvements in land cover classification. Technology demonstrations include a test of data compression on data quality, the first spaceborne utilization of short wave infrared Schottky barrier Pd2Si detector arrays, and the use of close-butted, multi-array modules with attached spectral filters.
Barnes, W. L.↗
Data processing 1: Advancements in machine analysis of multispectral data
Multispectral data processing procedures are outlined beginning with the data display process used to accomplish data editing and proceeding through clustering, feature selection criterion for error probability estimation, and sample clustering and sample classification. The effective utilization of large quantities of remote sensing data by formulating a three stage sampling model for evaluation of crop acreage estimates represents an improvement in determining the cost benefit relationship associated with remote sensing technology.
Swain, P. H.↗
Support Vector Machines for Classification of Direct Energy Deposition Standoff Distance for Improved Process Control
A critical factor in the implementation of direct energy deposition is the ability to maintain the standoff distance between the nozzle and the build surface, as this influences powder capture efficiency and overall part quality. Due to process-related variations, layer height may vary, causing unintended variation in standoff distance and poor build quality. While prior work has utilized contact probing to qualify standoff distance during processing, in situ methods for qualification of standoff distance are of major interest. The present work seeks to understand efficacy of image-based methods for classifying standoff distance variation in real-time using support vector machines (SVMs). It was hypothesized that the size of the melt pool and the amount of spatter will have significant correlations with deviations in the standoff distance; thus, SVMs were used on a dataset that is comprised of morphological features of melt pool size and image entropy. The SVM model was used to classify melt pool images into categories according to standoff distance variation from nominal. K-folds cross validation was used to find the optimal hyperparameters for the SVM model. To understand the impact of the selected features on the classification performance and inference speed, multiple models were trained with differing numbers of included features. Results for classification score, inference time, and image preprocessing/feature extraction from these data are reported. The present results show that the SVM model was able to predict the standoff distance classification with an accuracy of 97 percent and a speed of 0.122 s per image, making it a viable solution for real-time control of standoff distance.
Klesmith, Zoe↗
Interpretable machine learning models classify minerals via spectroscopy
Developing methods to identify mineral species confidently and rapidly from Raman spectral analysis is critical to numerous fields. Traditionally, analysis relies on pattern matching the Raman spectrum of an unknown dataset with a supporting library of well-characterized spectral data, which may prove difficult for environmental samples that are poorly crystalline or phase mixtures. Here, we developed interpretable machine learning models that can classify uranium minerals by secondary oxyanion chemistry and other physicochemical properties based solely on Raman spectra. This new ML method produces a mineral profile of physical and chemical properties for an unknown sample and can rapidly classify or identify unknown minerals from Raman data, without the need for an exact pattern match in a spectral library. Training models are validated by 1. Strong correlation of high confidence model regions with published spectroscopic assignments and 2. Correct classification of a mineral not present in training data. Training data are from the Compendium of Uranium Raman and Infrared Experimental Spectra and available crystallographic information files within the open-source Smart Spectral Matching scientific framework. Physically meaningful classifier models can rapidly identify key structural and chemical information about unknown uranium minerals and the overall methodology is broadly applicable for mineral phases.
Machine learning↗
Wheat cultivation: Identification and estimation of areas using LANDSAT data
The feasibility of using automatically processed multispectral data obtained from LANDSAT to identify wheat and estimate the areas planted with this grain was investigated. Three 20 km by 40 km segments in a wheat growing region of Rio Grande do Sul were aerially photographed using type 2443 Aerochrome film. Three maps corresponding to each segment were obtained from the analysis of the photographs which identified wheat, barley, fallow land, prepared soil, forests, and reforested land. Using basic information about the fields and maps made from the photographed areas, an automatic classification of wheat was made using MSS data from two different periods: July to September and July to October 1979. Results show that orbital data is not only useful in characterizing the growth of wheat, but also provides information of the intensity and extent of adverse climate which affects cultivation. The temporal and spatial characteristics of LANDSAR data are also demonstrated.
Dejesusparada, N.↗
ADP of multispectral scanner data for land use mapping
The advantages and disadvantages of various remote sensing instrumentation and analysis techniques are reviewed. The use of multispectral scanner data and the automatic data processing techniques are considered. A computer-aided analysis system for remote sensor data is described with emphasis on the image display, statistics processor, wavelength band selection, classification processor, and results display. Advanced techniques in using spectral and temporal data are also considered.
Hoffer, R. M.↗
A compositional classification scheme for meteoritic chondrules
A compositionally based classification scheme for chondrules is proposed that will help in systematizing the wealth of data available and disentangling the effects of nebular and subsequent processes. The classification is not by texture or the composition of a single phase, or a mixture of these two, but rather is a comprehensive, systematic approach which uses the composition of the two main chondrule components. This scheme is applicable to over 95 percent of the chondrules and is easily applied using an electron microprobe. It stresses the original diversity of the chondrules and the complex yet facile way in which they respond to parent-body metamorphism. Results using this classification scheme suggest that arguments against an important role of chondrules in determining the compositional trends of the chondrites have been premature.
Sears, Derek W. G.↗
Developing processing techniques for Skylab data
The author has identified the following significant results. The effects of misregistration and the scan-line-straightening algorithm on multispectral data were found to be: (1) there is greatly increased misregistration in scan-line-straightening data over conic data; (2) scanner caused misregistration between any pairs of channels may not be corrected for in scan-line-straightened data; and (3) this data will have few pure field center pixels than will conic data. A program SIMSIG was developed implementing the signature simulation model. Data processing stages of the experiment were carried out, and an analysis was made of the effects of spatial misregistration on field center classification accuracy. Fifteen signatures originally used for classifying the data were analyzed, showing the following breakdown: corn (4 signatures), trees (2), brush (1), grasses, weeds, etc. (5), bare soil (1), soybeans (1), and alfalfa (1).
Nalepka, R. F.↗
The preprocessing of multispectral data. II
It is pointed out that a correction of atmospheric effects is an important requirement for a full utilization of the possibilities provided by preprocessing techniques. The most significant characteristics of original and preprocessed data are considered, taking into account the solution of classification problems by means of the preprocessing procedure. Improvements obtainable with different preprocessing techniques are illustrated with the aid of examples involving Landsat data regarding an area in Colorado.
Quiel, F.↗
Taxonomy of asteroids
A taxonomic system for asteroids is discussed which is based on seven directly observable parameters from polarimetry, spectrophotometry, radiometry, and UBV photometry. The classification scheme is entirely empirical and independent of specific mineralogical interpretations. Five broad classes (designated C, S, M, E, and R), as well as an 'unclassifiable' designation, are defined on the basis of observational data for 523 asteroids. Computer-generated type classifications and derived diameters are given for the 523 asteroids, and the application of the classification procedure is illustrated. Of the 523 asteroids classified, 190 are identified as C objects, 141 as S type, 13 as type M, three as type E, three as type R, 55 as unclassifiable, and 118 as ambiguous. The present taxonomic system is compared with several other asteroid classification systems.
Bowell, E.↗
Levenberg–Marquardt multi-classification using hinge loss function
Incorporating higher-order optimization functions, such as Levenberg-Marquardt (LM) have revealed better generalizable solutions for deep learning problems. However, these higher-order optimization functions suffer from very large processing time and training complexity especially as training datasets become large, such as in multi-view classification problems, where finding global optima is a very costly problem. To solve this issue, we develop a solution for LM-enabled classification with, to the best of knowledge first-time implementation of hinge loss, for multiview classification. Hinge loss allows the neural network to converge faster and perform better than other loss functions such as logistic or square loss rates. Here we prove our method by experimenting with various multiclass classification challenges of varying complexity and training data size. The empirical results show the training time and accuracy rates achieved, highlighting how our method outperforms in all cases, especially when training time is limited. Our paper presents important results in the relationship between optimization and loss functions and how these can impact deep learning problems.
97 MATHEMATICS AND COMPUTING↗