Parallel Processing of Remotely Sensed Hyperspectral Images on Heterogeneous Clusters (original) (raw)
Related papers
International Journal of High Performance Computing Applications, 2008
The development of efficient techniques for transforming massive volumes of remotely sensed hyperspectral data into scientific understanding is critical for space-based Earth science and planetary exploration. Although most available parallel processing strategies for information extraction and mining from hyperspectral imagery assume homogeneity in the underlying computing platform, heterogeneous networks of computers (HNOCs) have become a promising cost-effective solution, expected to play a major role in many on-going and planned remote sensing missions. In this paper, we develop a new morphological parallel algorithm for hyperspectral image classification using HeteroMPI, an extension of MPI for programming high-performance computations on HNOCs. The main idea of HeteroMPI is to automate and optimize the selection of a group of processes that executes a heterogeneous algorithm faster than any other possible group in a heterogeneous environment. In order to analyze the impact of many-to-one (gather) communication operations introduced by our proposed algorithm, we resort to a recently proposed collective communication model. The parallel algorithm is validated using two heterogeneous clusters at University College Dublin and a massively parallel Beowulf cluster at NASA's Goddard Space Flight Center.
2006
Hyperspectral imaging is a new technique in remote sensing that generates hundreds of images corresponding to different wavelength channels for the same area on the surface of the Earth. Most available techniques for hyperspectral image classification focus on analyzing the data without incorporating the spatial information; i.e. the data is treated not as an image but as an unordered listing of spectral measurements where the spatial coordinates can be shuffled arbitrarily without affecting the final analysis. Despite the growing interest in the development of techniques for interpretation and classification of such high-dimensional imagery, only a few efforts devoted to the design of parallel implementations exist in the open literature. In this paper, we describe AMEEPAR, a parallel morphological algorithm that integrates the spatial and spectral information. The algorithm has been specifically optimized in this work for execution on heterogeneous networks of workstations. The parallel properties and classification accuracy of the proposed approach are evaluated using four networks of workstations distributed among different locations, and a massively parallel Beowulf cluster at NASA’s Goddard Space Flight Center.
Commodity cluster-based parallel processing of hyperspectral imagery
Journal of Parallel and Distributed Computing, 2006
The rapid development of space and computer technologies has made possible to store a large amount of remotely sensed image data, collected from heterogeneous sources. In particular, NASA is continuously gathering imagery data with hyperspectral Earth observing sensors such as the Airborne Visible-Infrared Imaging Spectrometer (AVIRIS) or the Hyperion imager aboard Earth Observing-1 (EO-1) spacecraft. The development of fast techniques for transforming the massive amount of collected data into scientific understanding is critical for spacebased Earth science and planetary exploration. This paper describes commodity cluster-based parallel data analysis strategies for hyperspectral imagery, a new class of image data that comprises hundreds of spectral bands at different wavelength channels for the same area on the surface of the Earth. An unsupervised technique that integrates the spatial and spectral information in the image data using multi-channel morphological transformations is parallelized and compared to other available parallel algorithms. The code's portability, reusability and scalability are illustrated by using two high-performance parallel computing architectures: a distributed memory, multiple instruction multiple data (MIMD)-style multicomputer at European Center for Parallelism of Barcelona, and a Beowulf cluster at NASA's Goddard Space Flight Center. Experimental results suggest that Beowulf clusters are a source of computational power that is both accessible and applicable to obtaining results in valid response times in information extraction applications from hyperspectral imagery. MD. His research interests are in remote sensing, digital image analysis, parallel and distributed computing, hardware-based architectures, operating systems management and configuration, and neural network-based pattern recognition.
2008
The wealth spatial and spectral information available from last-generation Earth observation instruments has introduced extremely high computational requirements in many applications. Most currently available parallel techniques treat remotely sensed data not as images, but as unordered listings of spectral measurements with no spatial arrangement. In thematic classification applications, however, the integration of spatial and spectral information can be greatly beneficial. Although such integrated approaches can be efficiently mapped in homogeneous commodity clusters, low-cost heterogeneous networks of computers (HNOCs) have soon become a standard tool of choice for dealing with the massive amount of image data produced by Earth observation missions. In this paper, we develop a new morphological/neural algorithm for parallel classification of high-dimensional (hyperspectral) remotely sensed image data sets. The algorithm's accuracy and parallel performance is tested in a variety of homogeneous and heterogeneous computing platforms, using two networks of workstations distributed among different locations, and also a massively parallel Beowulf cluster at NASA's Goddard Space Flight Center in Maryland.
2006
The development of efficient techniques for transforming the massive volume of remotely sensed hyperspectral data collected on a daily basis into scientific understanding is critical for space-based Earth science and planetary exploration. Although most available parallel processing strategies for hyperspectral image analysis assume homogeneity in the computing platform, heterogeneous networks of computers represent a promising cost-effective solution expected to play a major role in many on-going and planned remote sensing missions. To address the need for cost-effective parallel hyperspectral imaging algorithms, this paper develops an innovative heterogeneous parallel algorithm for spatial/spectral morphological analysis of hyperspectral image data. The algorithm has been developed using Heterogeneous MPI (HeteroMPI), an extension of MPI for programming high-performance computations on heterogeneous networks of computers. Experimental results are presented and discussed in the context of a realistic application, based on hyperspectral data collected by NASA's Jet Propulsion Laboratory.
Efficient information extraction from hyperspectral imagery using networks of workstations
2005
The rapid development in space and computer technologies has made possible to store a large amount of remotely sensed image data, collected from heterogeneous sources. In particular, NASA is continuously gathering imagery data with hyperspectral sensors such as the Airborne Visible-Infrared Imaging Spectrometer (AVIRIS) or the Hyperion imager aboard Earth Observing-1 (EO-1) spacecraft. The development of efficient techniques for transforming the massive amount of collected data into scientific understanding is critical for space-based Earth science and planetary exploration. Heterogeneous networks of workstations are a very promising cost-effective parallel computing architecture. Unlike traditional homogeneous parallel platforms, heterogeneous architectures are composed of processors running at different speeds. This heterogeneity results in distributed-memory parallel computing systems created from commodity components that can satisfy specific computational requirements for the Earth and space sciences community. This paper explores techniques for mapping hyperspectral image analysis algorithms onto heterogeneous networks of workstations. Important aspects in algorithm design such as portability, reusability and scalability are illustrated by using homogeneous and heterogeneous parallel computing facilities at NASA's Goddard Space Flight Center and, European Center for Parallelism of Barcelona, and University of Extremadura in Spain. Hyperspectral image data from the AVIRIS data repository is used in experiments, which reveal that heterogeneous networks of workstations are a source of computational power that is both accessible and applicable to obtaining results quickly enough for practical use in information extraction applications from hyperspectral imagery.
2006
Hyperspectral imaging is a new technique which has become increasingly important in many remote sensing applications, including automatic target recognition for military and defense/security deployment, risk/hazard prevention and response including wild land fire tracking, biological threat detection, monitoring of oil spills and other types of chemical contamination, etc. Hyperspectral imaging applications generate massive volumes of data and require timely responses for swift decisions which depend upon high computing performance of algorithm analysis. Although most currently available parallel processing strategies for hyperspectral image analysis assume homogeneity in the computing platform, heterogeneous networks of workstations represent a very promising cost-effective solution expected to play a major role in the design of highperformance computing platforms for many ongoing and planned remote sensing missions. This paper explores innovative techniques for mapping hyperspectral analysis algorithms onto heterogeneous networks of workstations available at NASA's Goddard Space Flight Center and University of Maryland. Experimental results reveal that heterogeneous networks of workstations represent a source of computational power that is both accessible and applicable in hyperspectral imaging studies.
Efficient Collective Communication Paradigms for Hyperspectral Imaging Algorithms Using HeteroMPI
2008
Most of the parallel strategies used for information extraction in remotely sensed hyperspectral imaging applications have been implemented in the form of parallel algorithms on both homogeneous and heterogeneous networks of computers. In this paper, we develop a study on efficient collective communications based on the usage of HeteroMPI for a parallel heterogeneous hyperspectral imaging algorithm which uses concepts of mathematical morphology.
2005
Hyperspectral sensors represent the most advanced instruments currently available for remote sensing of the Earth. The high spatial and spectral resolution of the images supplied by systems like the airborne visible infra-red imaging spectrometer (AVIRIS), developed by NASA Jet Propulsion Laboratory, allows their exploitation in diverse applications, such as detection and control of wild fires and hazardous agents in water and atmosphere, detection of military targets and management of natural resources. Even though the above applications require a response in real time, few solutions are available to provide fast and efficient analysis of these types of data. This is mainly caused by the dimensionality of hyperspectral images, which limits their exploitation in analysis scenarios where the spatial and temporal requirements are very high. In the present work, we describe a new parallel methodology which deals with most of the previously addressed problems. The computational performance of the proposed analysis methodology is evaluated using two parallel computer systems, a SGI Origin 2000 shared memory system located at the European Center of Parallelism of Barcelona, and the Thunderhead Beowulf cluster at NASA's Goddard Space Flight Center.
Parallel Computing, 2008
Imaging spectroscopy, also known as hyperspectral imaging, is a new technique that has gained tremendous popularity in many research areas, including satellite imaging and aerial reconnaissance. In particular, NASA is continuously gathering high-dimensional image data from the surface of the earth with hyperspectral sensors such as the Jet Propulsion Laboratory's Airborne Visible-Infrared Imaging Spectrometer (AVIRIS) or the Hyperion hyperspectral imager aboard NASA's Earth Observing-1 (EO-1) spacecraft. Despite the massive volume of scientific data commonly involved in hyperspectral imaging applications, very few parallel strategies for hyperspectral analysis are currently available, and most of them have been designed in the context of homogeneous computing platforms. However, heterogeneous networks of workstations represent a very promising cost-effective solution that is expected to play a major role in the design of high-performance computing platforms for many on-going and planned remote sensing missions. Our main goal in this paper is to understand parallel performance of hyperspectral imaging algorithms comprising the standard hyperspectral data processing chain (which includes pre-processing, selection of pure spectral components and linear spectral unmixing) in the context of fully heterogeneous computing platforms. For that purpose, we develop an exhaustive quantitative and comparative analysis of several available and new parallel hyperspectral imaging algorithms by comparing their efficiency on both a fully heterogeneous network of workstations and a massively parallel homogeneous cluster at NASA's Goddard Space Flight Center in Maryland.