John A. Rushing

Service-Oriented Environments for Dynamically Interacting with Mesoscale Weather

Within a decade after John von Neumann and colleagues conducted the first experimental weather forecast on the ENIAC computer in the late 1940s, numerical models of the atmosphere become the foundation of modern-day weather forecasting and one of the driving application areas in computer science. This article describes research that is enabling a major shift toward dynamically adaptive responses to rapidly changing environmental conditions.

ADaM: a data mining toolkit for scientists and engineers

Algorithm Development and Mining (ADaM) is a data mining toolkit designed for use with scientific data. It provides classification, clustering and association rule mining methods that are common to many data mining systems. In addition, it provides feature reduction capabilities, image processing, data cleaning and preprocessing capabilities that are of value when mining scientific data. The toolkit is packaged as a suite of independent components, which are designed to work in grid and cluster environments. The toolkit is extensible and scalable, and has been successfully used in several diverse data mining applications. ADaM has also been used in conjunction with other data mining toolkits and with point tools. This paper presents the architecture and design of the ADaM toolkit and discusses its application in detecting cumulus cloud fields in satellite imagery.

Techniques and Experience in Mining RemotelySensed Satellite Data

The paper presents a set of requirements for a datamining system for mining remotely sensed satellitedata based on a number of taxonomies that characterizemining of such data. The first of these taxonomies isbased on knowledge of the mining objectives and miningalgorithms. The second is based on variousrelationships that are found in data, including thosebetween different types of data, different spatiallocations of the data and different times of datacapture. The paper then describes the ADaM data miningsystem, which was developed to address theserequirements. The paper describes several data miningtechniques that have been applied to remotely senseddata. The first type is target independent mining,which mines data for transients and trends, with minedresults representing a highly concentrated form of theoriginal data. The second type is the mining ofvectors (representing multi-spectral or fused data)for association rules representing relationshipsbetween the various types of data represented by theelements of the vector. The third type mines data forassociation rules that characterize the texture of thedata.

Detection of cumulus cloud fields in satellite imagery

Boundary layer cumulus clouds are hard to detect in satellite imagery, especially for GOES imagery due to the coarse resolution of the IR channels. Two different approaches for the detection cumulus clouds in GOES satellite imagery are discussed and intercompared. The first step, structural thresholding, uses the morphology of cumulus cloud fields for detection. The second type, uses 1) classifiers based on texture and spectral, 2) edge detection and spectral, and 3) purely spectral features. For five selected scenes, cumulus cloud masks are created using these various methods and are compared against the expert-labeled masks. The structural thresholding method has the highest percentage of correct classification, followed by classifier based on Laplacian edge detection features. The classification time is lowest for the structural thresholding method, followed by classifiers based on spectral, edge detection, textural features. The structural thresholding method also is capable of detecting individual cumulus clouds within cloud fields. For the five scenes investigated, the average percentage of correct labeling of cumulus clouds by the structural thresholding method is 86 percent.

Real time target tracking with binary sensor networks and parallel computing

A parallel real time data fusion and target tracking algorithm for very large binary sensor networks is presented. A binary sensor can give an on or off signal to indicate the presence or absence of targets within its range, but it cannot tell how many targets are present, where the targets are, how fast they are moving, or which direction they are heading. In order to detect and track targets using these sensors, it is necessary to fuse information from more than one sensor. A parallel data fusion process based on simulated annealing is used to identify and locate targets. Processing is performed on a commodity Linux cluster with communication between nodes facilitated by the Message Passing Interface (MPI). The fusion and tracking algorithm is tested with a wide variety of sensor network parameters using target track data from a theater level air combat simulation. It is demonstrated that very accurate target detection and localization are possible even though the binary sensors themselves provide little information and have high error rates. Real time tracking is performed on a network with 2.5 million sensors on a commodity cluster with only 50 processors.

A Data Fusion Algorithm for Large Heterogeneous Sensor Networks

A distributed search based data fusion algorithm is presented for target detections in large heterogeneous sensor networks. A score function is introduced as the objection function during the optimal search. The network state is determined when the score is the highest. A close to optimal solution can be obtained before the arrival of the next sensor data thus enabling real time target tracking. The algorithm is evaluated with a series of real-time simulations on networks of variable sensor compositions with a commodity Linux cluster.

For scientific data discovery: why can't the archive be more like the Web?

The paper addresses the problem of acquiring from scientific data, metadata that is descriptive of the actual content of the data. Scientists can use this content based metadata in subsequent archive searches to find data sets of interest. Such metadata would be especially useful in large scientific archives such as NASAs Earth Observing System Data and Information System (EOSDIS). The paper presents two generic approaches for content based metadata acquisition: target dependent and target independent. Both of these approaches are oriented toward characterizing datasets in terms of the scientific phenomena, such as mesoscale convective systems (severe storms) that they contain. In the target dependent approach, the archived data is mined for particular phenomena of interest and polygons representing the phenomena are stored in a spatial database where they can be used in the data search process. In the target independent approach, data is initially mined for deviations from normal and for trends. This data can then be used for subsequent searches for particular transient phenomena using the deviation data, or for phenomena related to trends. The paper describes results from implementing both of these approaches.

Deep learning for multisensor image resolution enhancement

We describe a deep learning convolutional neural network (CNN) for enhancing low resolution multispectral satellite imagery without the use of a panchromatic image. For training, low resolution images are used as input and corresponding high resolution images are used as the target output (label). The CNN learns to automatically extract hierarchical features that can be used to enhance low resolution imagery. The trained network can then be effectively used for super-resolution enhancement of low resolution multispectral images where no corresponding high resolution image is available. The CNN enhances all four spectral bands of the low resolution image simultaneously and adjusts pixel values of the low resolution to match the dynamic range of the high resolution image. The CNN yields higher quality images than standard image resampling methods.

Visualizations for the spyglass ontology-based information analysis and retrieval system

Spyglass is an ontology-based information retrieval system designed to help analysts explore very large collections of unstructured text documents. The tool includes two main components: server and client. The server is a web-based service that uses a specific domain ontology to index a collection of documents, answer queries from the client, and provide retrieval and visualization services based on the ontology and the resulting index. The client is a graphical user interface which allows analysts to explore the document collections, query single or multiple entities of interest of the ontology and retrieve the documents relevant to the query. The rich set of visualization tools in Spyglass will be presented in this paper.

Real-time target tracking simulations in large disparate sensor networks

Real-time target tracking in large disparate sensor networks has been simulated with a parallelized search based data fusion algorithm using a simulated annealing approach. The networks are composed of large numbers of low fidelity binary and bearing-only sensors, and small numbers of high fidelity position sensors over a large region. The primitive sensors provide limited information, not sufficient to locate targets; the position sensors can report both range and direction of the targets. Target positions are determined through fusing information from all types of sensors. A score function, which takes into account the fidelity of sensors of different types, is defined and used as the evaluation function for the optimization search. The fusion algorithm is parallelized using spatial decomposition so that the fusion process can finish before the arrival of the next set of sensor data. A series of target tracking simulations are performed on a Linux cluster with communication between nodes facilitated by the Message Passing Interface (MPI). The probability of detection (POD), false alarm rate (FAR), and average deviation (AVD) are used to evaluate the network performance. The input target information used for all the simulations is a set of target track data created from a theater level air combat simulation.