Del Mar Photonics

Optics and Photonics Presentations attended by Del Mar Photonics team.

Hyperspectral imaging

Request a quote for hyperspectral optical components and systems

Hyperspectral Airborne Tactical Instrument (HATI): a low-cost compact airborne hyperspectral imager
Paper 7458-12 of Conference 7458
Date: Wednesday, 05 August 2009
Time: 2:00 PM – 2:20 PM

Author(s): Brian K. Baldauf, Stephanie R. Sandor-Leahy, Northrop Grumman Space Technology (United States)

 

The Mapping Reflectance Energy Sensor-MaRS: a new level of hyperspectral technology
Paper 7457-2 of Conference 7457
Date: Monday, 03 August 2009
Time: 10:00 AM – 10:30 AM

Author(s): Christopher Simi, Ernest Reith, National Geospatial-Intelligence Agency (United States); Robert Green, Michael Eastwood, Jet Propulsion Lab. (United States); Fred Olchowski, Logos Technologies, Inc. (United States)

 
The need to provide high-quality large-area hyperspectral data for land use, geological, environmental, and mapping applications is critical to the US Government in the 21st Century. Technological advances with regard to larger focal plane arrays (FPA) along with maturing spectrometer designs have made it possible for the development of a next generation system beyond AVIRIS. This paper will introduce the MaRS system along with some data examples from Cuprite, NV.

EnMAP hyperspectral imager: instrument design status, calibration, and operation approaches
Paper 7457-3 of Conference 7457
Date: Monday, 03 August 2009
Time: 10:30 AM – 10:50 AM

Author(s): Bernhard Sang, Stefan Hofer, Klaus-Peter Förster, Josef Schubert, Valery Mogulsky, Kayser-Threde GmbH (Germany); Hermann J. Kaufmann, GeoForschungsZentrum Potsdam e.V. (Germany); Andreas Neumann, Andreas Müller, Christian Chlebek, Thomas Eversberg, Deutsches Zentrum für Luft- und Raumfahrt e.V. (Germany)

 
The Environmental Mapping and Analysis Program (EnMAP) is a German space based hyperspectral mission with a targeted launch date in 2012. The single payload hyperspectral instrument uses a dual pushbroom configuration to image the ground at 30m GSD for wavelengths from 420nm to 2450nm with medium spectral resolution. The paper presents the design status of the instrument optical unit which consists of a thermally stabilized optomechanical system comprising two prism based spectrometers with their respective focal planes and a common telescope. Regular in-orbit calibration operations will allow correcting the sensor data to ensure high data consistency throughout the mission. The calibration concept and related on-board facilities will be discussed in conjunction with the sensor characteristics.

Hyperspectral-LIDAR system and data product integration for terrestrial applications
Paper 7457-4 of Conference 7457
Date: Monday, 03 August 2009
Time: 10:50 AM – 11:10 AM

Author(s): Lawrence A. Corp, Yen-Ben Cheng, Science Systems and Applications, Inc. (United States); Elizabeth M. Middleton, NASA Goddard Space Flight Ctr. (United States); Geoffrey H. Parker, Smithsonian Environmental Research Ctr. (United States); Karl F. Huemmrich, Petya K. E. Campbell, NASA Goddard Space Flight Ctr. (United States)

 
This manuscript details the development and validation of a unique forward thinking instrument and methodology for monitoring terrestrial carbon dynamics through synthesis of existing hyperspectal sensing and Light Detection and Ranging (LIDAR) technologies. The primary components of the Hyperspec-LIDAR system are the ruggedized Hyperspec VNIR Concentric Imaging Spectrometer and the LD90 single point laser range finder. The sensors are mounted on a heavy duty motorized pan-tilt which was programmed to support both push-broom style hyperspectral imaging and 3-D canopy LIDAR structural profiling. The goal of this research is to produce integrated science and data products from ground observations monitor Net Ecosystem Exchange.

Comparison of basis-vector selection methods for structural modeling of hyperspectral imagery
Paper 7457-29 of Conference 7457
Date: Monday, 03 August 2009
Time: 2:30 PM – 2:50 PM

Author(s): Carolina Peña-Ortega, Miguel Velez-Reyes, Univ. de Puerto Rico Mayagüez (United States)

 
This work presents a comparison of different methods for structural modeling of hyperspectral imagery for target detection. We study structured models based on linear subspaces and convex polyhedral cones and study how these structured models perform for target detection. Different training methods are studied: Singular Value Decomposition (SVD) is used for subspace modeling, and Maximum Distance (MaxD) and Positive Matrix Factorization (PMF) for convex polyhedral modeling. We study different detectors based on orthogonal and oblique projections for subspace and convex polyhedral cones and evaluate their performance using ROC curves. Experimental results using HYDICE imagery are presented.

A novel approach in endmember extraction in hyperspectral imagery
Paper 7457-11 of Conference 7457
Date: Monday, 03 August 2009
Time: 2:50 PM – 3:10 PM

Author(s): Yousef Rezaei, Mohammad Reza Mobasheri, Mohammad Javad Valadan Zouj, K.N.Toosi Univ. of Technology (Iran, Islamic Republic of)

 

Hyperspectral clustering for the study of Zapotec state formation
Paper 7457-12 of Conference 7457
Date: Monday, 03 August 2009
Time: 3:40 PM – 4:00 PM

Author(s): Justin D. Kwong, David W. Messinger, Rochester Institute of Technology (United States) and Digital Imaging and Remote Sensing Lab. (United States); William D. Middleton, Rochester Institute of Technology (United States)

 
This project will create full coverage classification maps of the Oaxaca, Mexico for use in studying Zapotec civilization. Hyperion imagery will be assessed using three clustering algorithms: IsoData, Gaussian Maximum Likelihood and Gradient Flow. The clustering process will be automated to efficiently classify approximately 30,000 km¬2 research area. Through analysis of algorithm parameters and region of interests, class maps with high-quality material resolution will be produced, differentiating classes into floral communities, urban expansion, etc. These class maps and other data in a GIS allow archaeologists to develop theories on topics such as settlement networks, resource availability, and environmental impact.

Spectral calibration in hyperspectral sounders
Paper 7452-16 of Conference 7452
Date: Monday, 03 August 2009
Time: 3:50 PM – 4:10 PM

Author(s): Evan M. Manning, Hartmut H. Aumann, Jet Propulsion Lab. (United States) and California Institute of Technology (United States); Larrabee Strow, Scott Hannon, Univ. of Maryland, Baltimore County (United States)

 
The knowledge of the frequencies of the spectral response functions (SRF) of the channels of hyperspectral sounders at the 10 ppmf level is adequate for the retrieval of temperature and moisture profiles and data assimilation for weather forecasting but not for climate applications. SI traceability and knowledge at the 1 ppmf level and better are required separate artifacts due to instrumental effects from effects due to climate change. We use examples from AIRS and IASI to discuss SI traceable spectral calibration using the upwelling radiance spectra which reach the 1 ppmf level.

Hyperspectral monitoring of chemically sensitive plant sentinels
Paper 7457-14 of Conference 7457
Date: Monday, 03 August 2009
Time: 4:20 PM – 4:40 PM

Author(s): Danielle A. Simmons, Nina G. Raqueno, John P. Kerekes, Rochester Institute of Technology (United States)

 
Plants can be effective bio-sensors for the automated detection of a chemical threat. Monitoring the bio-sensors requires a specifically tailored hyperspectral system. Tobacco plants were genetically engineered to de-green when exposed to a chemical of interest. The reflectance spectra of the bio-sensors were characterized during the de-greening process in a laboratory environment. RIT’s DIRSIG was used to simulate detection of de-greened plants in the field. Trade studies of the bio-sensor monitoring system were also conducted using DIRSIG simulations. Preliminary results show that the most significant change in reflectance during the degreening period occurs in the near infrared region.

Vegetation water content at 970 nm: estimation using hyperspectral vegetation indices
Paper 7457-15 of Conference 7457
Date: Monday, 03 August 2009
Time: 4:40 PM – 5:00 PM

Author(s): Eric A. Salas, Geoffrey M. Henebry, South Dakota State Univ. (United States)

 
This study takes advantage of the water absorption feature at 970 nm that is visible on the vegetation spectrum using hyperspectral field data. The downward deflection at 970 nm is not very pronounced but visibly recognizable and can be used to estimate remotely the canopy water content. The individual performance of a number of existing vegetation water content (VWC) indices against the 970 nm absorption are assessed using linear regression model to establish relationships. The new Combined Vegetation Water Index (CVWI) developed showed a promise in assessing the vegetation water status derived from the 970 nm absorption wavelength.

Underwater unmixing and water optical properties retrieval using HyCIAT
Paper 7457-30 of Conference 7457
Date: Monday, 03 August 2009
Time: 5:00 PM – 5:20 PM

Author(s): Maria C. Torres-Madronero, Miguel Velez-Reyes, James A. Goodman, Univ. de Puerto Rico Mayagüez (United States)

 
Unmixing of the bottom for benthic habitat mapping in shallow coastal waters is a difficult problem due to the confounding effects of space variant bathymetry and water optical properties that result in signatures for the same benthic bottom type to look different at the water surface. This paper discusses different approaches to modify the linear unmixing approach to account for variable water optical properties and bathymetry and their implementation in the Hyperspectral Coastal Image Analysis Toolbox, HyCIAT. This toolbox allows the processing of hyperspectral image of shallow coastal areas to estimate water column optical properties, bathymetry, and perform unmixing for bottom composition. HyCIAT has been developed as part of the UPRM Solutionware project to develop software tools for hyperspectral image processing. The tool has been developed under the MATLAB environment and includes a series of algorithms developed by UPRM researchers under a graphical user interface that facilitates its use by the remote sensing community. The paper describes algorithms implemented in the toolbox, gives an overview of the graphical user interface, and presents results of its application to AVIRIS and AISA hyperspectral imagery collected over Kaneohe Bay in Hawaii and over Southwestern Puerto Rico.

Request a quote for hyperspectral optical components and systems

Hyperspectral image compression using low complexity integer KLT and three-dimensional asymmetric significance tree
Paper 7444A-17 of Conference 7444A
Date: Monday, 03 August 2009
Time: 6:00 PM

Author(s): Jing Huang, Nanjing Univ. of Posts and Telecommunications (China); Ri-hong Zhu, Nanjing Univ. of Science & Technology (China)

 
A lossy to lossless three-dimensional (3D) compression of hyperspectral imagery is presented. On the spectral dimension, a low complexity reversible integer Karhunen-Loève transform (KLT) is used to fully exploit the spectral redundancy, while two-dimensional integer wavelet transform is applied on the spatial dimension. Each coefficients are then encoded by the significance test of the 3D asymmetric significance tree node at each bitplane instead of ordered lists to track the significance status of the tree or block sets and coefficients. This algorithm has low complexity and can be applied to lossy to lossless progressive transmission.

Smart hyperspectral imaging detection based on electrically tunable liquid crystal Fabry-Perot microstructure array
Paper 7457-26 of Conference 7457
Date: Monday, 03 August 2009
Time: 6:00 PM

Author(s): Kan Liu, Dehua Li, Hui Li, XinYu Zhang, Tianxu Zhang, Huazhong Univ. of Science and Technology (China)

 
A smart spectral imaging detection method based on the integration of electrically tunable liquid-crystal Fabry-Perot microstructure array is proposed. It has wide application in many fields with advantages of low cost, highly compact integration. The structure can theoretically get hundreds of spectral bands simultaneously in one frame of the picture. Analysis on some key issues of liquid-crystal Fabry-Perot structures for imaging application and calibration problem by liquid crystal lens are made. Prototypic devices of 4×4 array with the cavity thickness ranging from 4~20 μm for the wavelength in the range of 800~900nm, are made and tested.

Hyperspectral information compression
Paper 7455-1 of Conference 7455
Date: Tuesday, 04 August 2009
Time: 8:10 AM – 8:40 AM

Author(s): Chein-I Chang, Univ. of Maryland, Baltimore County (United States)

 
Hyperspectral compression has received considerable interest in recent years. In order to measure the effectiveness of compression a common criterion is the compression ratio which is used to indicate how much data size is reduced after compression. In this case, we actually perform data compression. While the data size compression is of major interest in hyperspectral compression, a more crucial and critical issue encountered in hyperspectral compression is information preservation which is more important than data size compression. In order to address this need, this paper investigates a new concept, called Hyperpsectral Information Compression (HIC) as opposed to Hyperspectral Data Compression (HDC) commonly used in the literature.

Hyperspectral image lossless compression algorithm based on AP adaptive band regrouping
Paper 7455-3 of Conference 7455
Date: Tuesday, 04 August 2009
Time: 9:00 AM – 9:20 AM

Author(s): Mingyi He, Lin Bai, Yuchao Dai, Jing Zhang, Northwestern Polytechnical Univ. (China)

 
To obtain better compression performance, redundancy information has been well utilized for hyperspectral image compression. In this paper, a new algorithm for lossless compression of hyperspectral images is proposed. At first, an adaptive band regrouping scheme is proposed to regroup spectral images according to their statistic correlation. Secondly, the Affinity Propagation (AP) clustering approach is employed for the adaptive regrouping of bands as its rapid convergence in clustering. Thirdly, a linear context prediction is applied on the hyperspectral images in different groups. As comparisons, the experiments with 224 band AVIRIS hyperspectral images by using the new method and a few update methods including PDPCM, 3DSPIHT, CCSDS, JPEG2000 and ICIER were conducted. The compression ratio is respectively 3.46, 4.40, 4.43, 4.58, 4.49, 4.40 bpp. This result shows that the proposed method obtains a better compression performance with the other 5 comparison algorithms.

A portable solid-state high-spectral resolution hyperspectral imager
Paper 7457-20 of Conference 7457
Date: Tuesday, 04 August 2009
Time: 9:30 AM – 9:50 AM

Author(s): John Noto, Steven R. Watchorn, Robert B. Kerr, Scientific Solutions, Inc. (United States)

 
An imager based upon an etched liquid-crystal Fabry-Perot (LCFP) dispersive element can simultaneously sample four distinct resolution elements in the region 800 nm – 1100 nm, and tune in milliseconds to any one of 194,580 possible four-color scene images with a spectral resolution of 0.625 nm. Independently tunable quadrants of a single LCFP etalon are created by etching a transparent conducting layer on the etalon substrate, and one image from each quadrant is formed on a focal-plane array detector. Weighing less than 15 lbs. the portable, solid-state camera system is designed to provide fast RGB images of transient spectral phenomena.

Hyperspectral pixels in 2D imaging FPAs
Paper 7467-20 of Conference 7467
Date: Tuesday, 04 August 2009
Time: 10:20 AM – 10:40 AM

Author(s): Paul D. LeVan, Air Force Research Lab. (United States); Brian P. Beecken, Bethel Univ. (United States)

 
Exploiting the "third dimension" of dual-waveband FPAs, different grating orders have been paired with dualband wavelengths, allowing high efficiency hyperspectral imaging over broad wavelength regions. As time progresses, multi-waveband FPAs will provide an increase in spectral information at the pixel level. As the number of wavebands increases to the point of providing spectral overlap of adjacent spectral resolution elements, hyperspectral capability is then achieved by the FPA acting alone (without the need for external optical elements). This approach may become possible through advanced architectures, with photons of different wavelength absorbed at different depths, and photocurrents isolated with a vertical grid of contacts.

Bit allocation for 2D compression of hyperspectral images for classification
Paper 7455-6 of Conference 7455
Date: Tuesday, 04 August 2009
Time: 10:30 AM – 10:50 AM

Author(s): Sangwook Lee, Chulhee Lee, Yonsei Univ. (Korea, Republic of)

 
In this paper, we propose a bit allocation strategy for 2D compression methods for hyperspectral images. First, we select a number of classes from the original hyperspectral images. It is noted that the classes can be automatically selected by applying an unsupervised segmentation method. Then, we apply a feature extraction method and determine discriminantly dominant feature vectors. By examining the feature vectors, we determine the discriminant usefulness of each spectral band. Finally, based on the discriminant usefulness of spectral bands, we determine the allocation of each spectral band. Experimental results show that it is possible to enhance the discriminant information at the expense of PSNR. Depending on applications, one can either minimize the mean squared error or choose to preserve the classification capability of the hyperspectral images.

Progressive dimensionality reduction for hyperspectral imagery
Paper 7455-7 of Conference 7455
Date: Tuesday, 04 August 2009
Time: 10:50 AM – 11:10 AM

Author(s): Haleh Safavi, Chein-I Chang, Univ. of Maryland, Baltimore County (United States)

 
This paper attempts to address both issues by developing a new concept, called Progressive Dimensionality Reduction (PDR) which can perform data dimensionality progressive in terms of information preservation. Two procedures can be used to perform PDR in a forward or backward manner, referred to forward PDR (FPDR) or backward PDR (BPDR) where FPDR starts with a minimum number of spectral-transformed dimensions and increases the spectral-transformed dimension progressively as opposed to BPDR begins with a maximum number of spectral-transformed dimensions and decreases the spectral-transformed dimension progressively. Both procedures are terminated when a stopping rule is satisfied.
 

Request a quote for hyperspectral optical components and systems


Hyperspectral detection algorithms: use covariances or subspaces?
Paper 7457-22 of Conference 7457
Date: Tuesday, 04 August 2009
Time: 11:00 AM – 11:20 AM

Author(s): Dimitris G. Manolakis, MIT Lincoln Lab. (United States); Ronald B. Lockwood, Thomas W. Cooley, John Jacobson, Air Force Research Lab. (United States)

 
There are two broad classes of hyperspectral detection algorithm: Algorithms in the first class use the spectral covariance matrix of the background clutter; in contrast, algorithms in the second class characterize the background using a subspace model. In this paper we show that, due to the nature of hyperspectral imaging data, the two families of algorithms are intimately related. The link between the two representations of the background clutter is the low-rank of the covariance matrix of natural hyperspectral backgrounds and its relation to the spectral linear mixture model.

Soft-decision hyperspectral measures for target discrimination and classification
Paper 7457-24 of Conference 7457
Date: Tuesday, 04 August 2009
Time: 11:40 AM – 12:00 PM

Author(s): Chao-Cheng Wu, Chein-I Chang, Univ. of Maryland, Baltimore County (United States)

 
This paper develops a new class of hyperspectral measures, called soft-decision hyperspectral measures which use the similarity value between a data sample and a target signature or a class as an indicator of the likelihood of the data sample assigned to this particular signature or class instead of signature or class assignment which is a deterministic decision. In order for a soft-decision to perform target discrimination or classification, the soft-decision hyperspectral measure-generated likelihood values are normalized to probabilities so that a threshold can be used to make hard decisions via a recently developed 3D ROC analysis.

A neural network approach for improved detector performance of spectral matched filters in hyperspectral imagery
Paper 7457-25 of Conference 7457
Date: Tuesday, 04 August 2009
Time: 12:00 PM – 12:20 PM

Author(s): Robert S. Rand, National Geospatial-Intelligence Agency (United States)

 
The detection of subpixel materials in a hyperspectral scene is often accomplished using spectral matched filters or subspace projection. These methods rely on estimates of the background statistics or subspaces in a scene. A number of spectral matched filter methods have been developed with increasing sophistication, but recent research indicates that the method used to compute the background statistics may have a greater impact on overall detector performance. This research investigates the use of a neural network approach to estimate the background statistics needed for commonly used spectral matched filters.

Classification of hyperspectral colon tissue samples using MCMC with wavelet transforms
Paper 7446-43 of Conference 7446
Date: Tuesday, 04 August 2009
Time: 1:30 PM – 1:50 PM

Author(s): Khalid Masood, Nasir M. Rajpoot, Irfan T. Butt, The Univ. of Warwick (United Kingdom); Muhammad Yousaf, The Univ. of Texas at Arlington (United States)

 
In this paper we focus on the classification of hyperspectral colon biopsy samples using wavelet packets and MCMC simulations. The main objective is to model the shape descriptors of the tissue glands. Our approach is based on the idea that development of the colon cancer alters the macroarchitecture of the tissue glands. The algorithm is implemented in three phases. In phase I, dimensionality reduction and segmentation of the hyperspectral data is performed. Wavelet packets and Daubechies-4 filters are used to reduce the dimensionality of hyperspectral data and segment the tissue in three regions. Shape modelling using MCMC is followed after the segmentation. Finally, in phase III, moments invariants and Zernike moments are used for the classification. Experimental results indicate that the incorporation of wavelet packets with shape descriptors labels the two classes with high accuracy.

Virtual dimensionality for hyperspectral imagery (Plenary)
Paper 7461-501 of Conference 7461
Date: Tuesday, 04 August 2009

Author(s): Chein-I Chang, Univ. of Maryland, Baltimore County (United States)

Show Abstract
Virtual Dimensionality (VD) was recently developed to address the issue of how many spectrally distinct signatures in hyperspectral data. It originates from the pigeon-hole principle where each pigeon-hole is supposed to accommodate a pigeon. If a signal source is interpreted as a pigeon and a band as a pigeon-hole, then a spectral channel/band can be used to accommodate one signal source. To materialize this idea it is formatted as a binary composite hypothesis testing problem where the Neyman-Pearson detector is designed to determine the VD. This talk discusses the design rationale of VD and its utility in hyperspectral analysis.

Virtual dimensionality for hyperspectral imagery (Plenary)
Paper 7458-501 of Conference 7458
Date: Tuesday, 04 August 2009

Author(s): Chein-I Chang, Univ. of Maryland, Baltimore County (United States)

Show Abstract
Virtual Dimensionality (VD) was recently developed to address the issue of how many spectrally distinct signatures in hyperspectral data. It originates from the pigeon-hole principle where each pigeon-hole is supposed to accommodate a pigeon. If a signal source is interpreted as a pigeon and a band as a pigeon-hole, then a spectral channel/band can be used to accommodate one signal source. To materialize this idea it is formatted as a binary composite hypothesis testing problem where the Neyman-Pearson detector is designed to determine the VD. This talk discusses the design rationale of VD and its utility in hyperspectral analysis.

Virtual dimensionality for hyperspectral imagery (Plenary)
Paper 7457-501 of Conference 7457
Date: Tuesday, 04 August 2009

Author(s): Chein-I Chang, Univ. of Maryland, Baltimore County (United States)

Show Abstract
Virtual Dimensionality (VD) was recently developed to address the issue of how many spectrally distinct signatures in hyperspectral data. It originates from the pigeon-hole principle where each pigeon-hole is supposed to accommodate a pigeon. If a signal source is interpreted as a pigeon and a band as a pigeon-hole, then a spectral channel/band can be used to accommodate one signal source. To materialize this idea it is formatted as a binary composite hypothesis testing problem where the Neyman-Pearson detector is designed to determine the VD. This talk discusses the design rationale of VD and its utility in hyperspectral analysis.

Virtual dimensionality for hyperspectral imagery (Plenary)
Paper 7455-501 of Conference 7455
Date: Tuesday, 04 August 2009

Author(s): Chein-I Chang, Univ. of Maryland, Baltimore County (United States)

Show Abstract
Virtual Dimensionality (VD) was recently developed to address the issue of how many spectrally distinct signatures in hyperspectral data. It originates from the pigeon-hole principle where each pigeon-hole is supposed to accommodate a pigeon. If a signal source is interpreted as a pigeon and a band as a pigeon-hole, then a spectral channel/band can be used to accommodate one signal source. To materialize this idea it is formatted as a binary composite hypothesis testing problem where the Neyman-Pearson detector is designed to determine the VD. This talk discusses the design rationale of VD and its utility in hyperspectral analysis.



Virtual dimensionality for hyperspectral imagery (Plenary)
Paper 7452-501 of Conference 7452
Date: Tuesday, 04 August 2009

Author(s): Chein-I Chang, Univ. of Maryland, Baltimore County (United States)

Show Abstract
Virtual Dimensionality (VD) was recently developed to address the issue of how many spectrally distinct signatures in hyperspectral data. It originates from the pigeon-hole principle where each pigeon-hole is supposed to accommodate a pigeon. If a signal source is interpreted as a pigeon and a band as a pigeon-hole, then a spectral channel/band can be used to accommodate one signal source. To materialize this idea it is formatted as a binary composite hypothesis testing problem where the Neyman-Pearson detector is designed to determine the VD. This talk discusses the design rationale of VD and its utility in hyperspectral analysis.

Radiometric characterization of hyperspectral imagers using multispectral sensors
Paper 7452-38 of Conference 7452
Date: Tuesday, 04 August 2009
Time: 4:50 PM – 5:10 PM

Author(s): Joel T. McCorkel, College of Optical Sciences, The Univ. of Arizona (United States); Kurtis J. Thome, NASA Goddard Space Flight Ctr. (United States); Nathan P. Leisso, Nikolaus Anderson, Jeffrey S. Czapla-Myers, College of Optical Sciences, The Univ. of Arizona (United States)

 
The Remote Sensing Group (RSG) at the University of Arizona has a long history of using ground-based test sites for the calibration of airborne and satellite based sensors. Often, ground-truth measurements at these tests sites are not always successful due to weather and funding availability. Therefore, RSG has also employed automated ground instrument approaches and cross-calibration methods to verify radiometric calibration of a sensor. This work relies on cross-calibration methods using the Moderate Resolution Imaging Spectroradiometer (MODIS) as a reference for the hyperspectral sensor Hyperion. Hyperion has 220 spectral bands while MODIS has 32 spectral bands, nine of which are within the spectral and gain region covered by Hyperion. Hyperspectral reflectance derived from multispectral sensor data and prior site knowledge, along with automated atmospheric data products, is used in the radiometric characterization of a hyperspectral imager.

Lossy hyperspectral image compression tuned for spectral mixture analysis applications on NVidia graphics processing units
Paper 7458-14 of Conference 7458
Date: Tuesday, 04 August 2009
Time: 5:20 PM – 5:40 PM

Author(s): Antonio J. Plaza, Javier Plaza, Sergio Sanchez, Abel Paz, Univ. de Extremadura (Spain)

 
We develop a computationally efficient approach for lossy compression of hyperspectral scenes which has been specifically tuned to preserve the relevant information required in spectral mixture analysis (SMA) applications. The proposed method has been implemented in graphics processing units (GPUs) of Nvidia type. The proposed method can achieve very high compression ratios when applied to standard hyperspectral data sets, and can also retain the relevant information required for spectral unmixing in a computationally efficient way, achieving speedups in the order of 30 on a NVidia GeForce 8800 GTX graphic card when compared to an optimized implementation of the same code in a dual-core CPU.

Lossy hyperspectral image compression tuned for spectral mixture analysis applications on NVidia graphics processing units
Paper 7455-14 of Conference 7455
Date: Tuesday, 04 August 2009
Time: 5:20 PM – 5:40 PM

Author(s): Antonio J. Plaza, Javier Plaza, Sergio Sanchez, Abel Paz, Univ. de Extremadura (Spain)

 
We develop a computationally efficient approach for lossy compression of hyperspectral scenes which has been specifically tuned to preserve the relevant information required in spectral mixture analysis (SMA) applications. The proposed method has been implemented in graphics processing units (GPUs) of Nvidia type. The proposed method can achieve very high compression ratios when applied to standard hyperspectral data sets, and can also retain the relevant information required for spectral unmixing in a computationally efficient way, achieving speedups in the order of 30 on a NVidia GeForce 8800 GTX graphic card when compared to an optimized implementation of the same code in a dual-core CPU.

Hyperspectral two-photon near-infrared cancer imaging in-vitro and in-vivo
Paper 7413-26 of Conference 7413
Date: Tuesday, 04 August 2009
Time: 8:00 PM

Author(s): Nikolay S. Makarov, Jean R. Starkey, Mikhail A. Drobizhev, Aleksander K. Rebane, Montana State Univ., Bozeman (United States)

 
Two-photon excited near-infrared fluorescence allows for deeper penetration in biological tissues providing a unique tool for non-invasive detection and diagnostics of cancer.
We present a novel way of deep cancer detection based on multi-wavelength two-photon excited fluorescence of Styryl-9M dye. We show that combining the excitation at 1100-1200 nm with detection of fluorescence at 800 nm provides for minimal scattering and background absorption in a biological tissue.
We show that the ratio of fluorescence signals excited at 1100 and 1200 nm is sensitive to composition of the phantoms, and differentiates between those with normal or cancer cells. The method is a promising tool for non-invasive deep photodetection of cancer.

Course: Multispectral and Hyperspectral Image Sensors
Date: Wednesday, 05 August 2009
Time: 8:30 AM – 12:30 PM

Instructor(s): Terrence S. Lomheim, The Aerospace Corp. (United States)

View Full Details Add to My Schedule

Segmented PCA and JPEG2000 for hyperspectral image compression
Paper 7455-17 of Conference 7455
Date: Wednesday, 05 August 2009
Time: 8:40 AM – 9:00 AM

Author(s): Qian Du, Wei Zhu, Mississippi State Univ. (United States)

 
Principal component analysis (PCA) can outperform in spectral coding and a superior rate distortion performance can be provided in conjunction with JPEG2000 spatial coding. However, there is a minimum bitrate for it to perform appropriately. This is due to the fact that a large spectral transform matrix needs to be transmitted. This bitrate is positively proportional to the number of bands and inversely proportional to the number of pixels. This means for a hyperspectral data with a smaller spatial size, it may not be a good choice. We propose to replace PCA with segmented-PCA, which has lower computational complexity. It keeps the ratio of spatial/spectral size in a reasonable level so that PCA+JPEG2000 can work well at low bitrates. Even for images with large spatial size, segmented-PCA can improve compression performance because: 1) the overhead is reduced; 2) PCA is applied to the bands with higher spectral correlation. They have comparable impact on detection and classification.

Optimal wavebands searching prediction for lossless compression of hyperspectral imagery
Paper 7455-19 of Conference 7455
Date: Wednesday, 05 August 2009
Time: 9:20 AM – 9:40 AM

Author(s): Lang Wang, Shuxu Guo, Jilin Univ. (China)

 
The paper presents a new lossless compression algorithm for hyperspectral image, which using prediction tree method along with couple wavebands prediction. The algorithm first forms a four side neighborhood prediction tree and utilizes the previous encoded pixel to estimate the current one from the tree. Then we built an adaptive waveband selection mechanism of binary tree model to searching optimal couple prediction waveband for exploiting hyperspectral image’s spectrum correlation. Thirdly, bit-plane coding is adopted to compress the residuals. Experiments result shows that the approach proposed outperforms on compression ratio with higher average PSNR and low computational complexity.

Design and analysis of real-time endmember extraction algorithms for hyperspectral imagery
Paper 7455-21 of Conference 7455
Date: Wednesday, 05 August 2009
Time: 10:30 AM – 10:50 AM

Author(s): Chein-I Chang, Chao-Cheng Wu, Univ. of Maryland, Baltimore County (United States)

 
Endmember extraction has recently received considerable attention in hyperspectral data exploitation since they represent crucial and vital information for hyperspectral data analysis. An endmember is defined as an idealized signature and may or may not exist as a data sample or an image pixel. The interest of endmember extraction arises in the use of hundreds of contiguous spectral channels that allows a hyperspectral imaging sensor to uncover many subtle substances in diagnostic bands. However finding such substances also presents a great challenge to hyperspectral data analysts. In order to address this need many endmember extraction algorithms have been developed and designed in the past, but no work has been reported on how to implement endmember extraction algorithms in real-time. This paper investigates this issue in designing algorithms for real time processing of endmember extraction and developed several endmember extraction algorithms that can be implemented in real time.

Request a quote for hyperspectral optical components and systems

Massively parallel processing of remotely sensed hyperspectral images
Paper 7455-23 of Conference 7455
Date: Wednesday, 05 August 2009
Time: 11:10 AM – 11:30 AM

Author(s): Javier Plaza, Antonio J. Plaza, Abel Paz, Univ. de Extremadura (Spain)

 
Hyperspectral imaging (also known as imaging spectroscopy) is an emerging technique that has gained tremendous popularity in many research areas, most notably, in remotely sensed satellite imaging and aerial reconnaissance. In this paper, we develop several new parallel techniques for hyperspectral image processing that have been specifically designed to be run on massively parallel systems. The techniques developed cover the three following areas: 1) spectral mixture analysis, a popular approach to characterize mixed pixels in hyperspectral data; 2) unsupervised and supervised classification; and 3) lossy hyperspectral data compression. The scalability of these parallel techniques to a very high number of processors is investigated using the Discover cluster with 4,648 processors and 6712 GB of memory at NASA's Goddard Space Flight Center in Greenbelt, Maryland.

An ad-hoc approach for quality assessment of hyperspectral datacubes in target detection
Paper 7455-31 of Conference 7455
Date: Wednesday, 05 August 2009
Time: 4:00 PM – 4:20 PM

Author(s): Reza Rashidi Far, Shen-en Qian, Canadian Space Agency (Canada)

 
The problem of assessing different processing techniques for hyperspectral images target detection is addressed in this paper. An ad-hoc quality assessment approach is adopted in this paper to compare different noise reduction techniques of hyperspectral images for target detection applications. Two different noise reduction techniques are applied to a datacube collected over a well-studied area with human made targets. The quality of these noise reduced datacubes in preserving the identity of the targets of interest is compared with that of the original datacube. This is achieved by applying different measures on the datacubes. First, the Virtual Dimensionality (VD) is tried and the results for both of the noise reduction methods are compared with those of the original datacube for different false-alarm probability. Then Maximum Noise Fraction MNF) is applied to the datacubes and its capability in finding a rotation in which the information of the datacube is represented in a smaller number of bands is assessed. Finally using set measures and knowing the location of the targets, different classes are defined and the interclass and intra-class distances for each datacube is measured.

Automated display of hyperspectral images with unsupervised segmentation
Paper 7455-33 of Conference 7455
Date: Wednesday, 05 August 2009
Time: 4:40 PM – 5:00 PM

Author(s): Sangwook Lee, Chulhee Lee, Yonsei Univ. (Korea, Republic of)

 
In this paper, we investigate automated display methods for hyperspectral images. First, we apply an unsupervised segmentation method, which will produce a number of unlabeled classes. Then, we choose the classes whose sizes are larger than a threshold value. Then, we apply a feature extraction method to the chosen classes and find dominant features, which are used to display the hyperspectral images. We also exploit the use of the principal component analysis for the display of hyperspectral images. Experimental images show that the color images produced by the proposed methods show interesting characteristics compared to the conventional pseudo-color image.

Hyperspectral image compression based on the framework of DSC using 3D-wavelet and LDPC
Paper 7455-38 of Conference 7455
Date: Wednesday, 05 August 2009
Time: 5:30 PM

Author(s): Jiaji Wu, Kun Jiang, Xidian Univ. (China); Yong Fang, Northwest A&F Univ. (China)

 
Hyperspectral images include both spatial and spectral redundancies. In this paper, we propose a method based on both asymmetric 3D-wavelet transform and LDPC (low-density parity-check codes) to realize the compression of hyperspectral images in the framework of DSC (distributed source coding). Our methods can get rid of the correlation in a lower encoding complexity and get good performance for compression. Experimental results show, the SNR of our proposed algorithm is about 0.2-0.4dB higher than the performance of 3D-SPIHT. Compared with 2D-SPIHT and JPEG2000, The new method can obtain about 3-5dB gain.

Object detection in hyperspectral imagery using normalized cross-spectrum energy
Paper 7442-57 of Conference 7442
Date: Wednesday, 05 August 2009
Time: 5:30 PM

Author(s): Mohamed I. Elbakary, Mohammad S. Alam, Univ. of South Alabama (United States)

 
Hyperspectral sensors can facilitate automatic pattern recognition in cluttered imagery since manmade objects often differ considerably from the natural background in absorbing and reflecting the radiation at various wavelengths i.e., the identification of the objects is based on spectral signature of the objects in the scene. Normalized cross spectrum (cross-phase spectrum) has been extensively used for image registration. In this paper, we introduce preliminary results for a new approach for object detection in hyperspectral imagery by employing normalized cross spectrum. Normalized cross spectrum is employed as similarity measure between the spectral signature of known object and the investigated spectral signatures in the data. The new algorithm uses the advantages of the shape of the peak of the correlation to detect the pattern of interest. The proposed algorithm has been tested using real life hyperspectral imagery and the results show the effectiveness of the proposed approach.

Pattern Recognition in Hyperspectral Imagery Using Spectral JTC (Invited Paper)
Paper 7442-25 of Conference 7442
Date: Thursday, 06 August 2009
Time: 8:30 AM – 8:50 AM

Author(s): Mohammad S. Alam, Univ. of South Alabama (United States)

 
Pattern recognition in Hyperspectral imagery is a challenging problem due to the minute nature of the target signature and the requirement to process huge amount of data. In this paper, we investigate the recent trends and advancements in joint transform correlation (JTC) based pattern recognition in hyperspectral imagery. In particular, we investigate the application of spectral fringe-adjusted JTC for efficient target recognition in hyperspectral imagery. Techniques for eliminating false target detection, minimizing effects of noise and other artifacts will be considered. The performance of the spectral fringe-adjusted JTC will be compared with existing techniques by generating ROC curves using real life hyperspectral datasets.

An anomaly correlation skill score for the evaluation of the performance of hyperspectral sounders
Paper 7456-33 of Conference 7456
Date: Thursday, 06 August 2009
Time: 11:10 AM – 11:30 AM

Author(s): Hartmut H. Aumann, Evan M. Manning, Jet Propulsion Lab. (United States); Christopher D. Barnet, Eric S. Maddy, National Oceanic and Atmospheric Administration (United States)

 
The accuracy of temperature and water vapor profile retrievals is commonly discussed in terms of the rms error relative to truth data. With the availability of very accurate forecasts and the increasing use of the forecast for the initialization of retrievals, an accuracy metric alone tends to produce misleading results. A much more relevant characterization is retrieval skill, which we define as the ability of an algorithm to get closer to the truth than the forecast, when the truth differs significantly from the forecast, i.e. the anomaly correlation between the retrieved and forecast versus truth and forecast. We explore retrieval skill with a number of different retrievals of temperature and water vapor profiles over land and ocean using Atmospheric Infrared Sounder (AIRS) and Infrared Atmospheric Sounding Interferometer (IASI) infrared hyperspectral sounder data.

Comparison of forest LAI estimated from LiDAR and hyperspectral data
Paper 7454-25 of Conference 7454
Date: Thursday, 06 August 2009
Time: 11:20 AM – 11:40 AM

Author(s): Yong Pang, Chinese Academy of Forestry (China)

 
This study select coniferous forests site and organized field measurement and airborne data collection campaign in June of 2008. It compared the performances of LAI estimation using LiDAR and hyperspectral data. The preliminary result shows both LiDAR and hyperspectral data were strongly correlated to field measured LAI, and hence, both data types are suitable for large scale mapping of LAI in forests. As self-contained accurate terrain information, the LiDAR data need less calibration samples from field measurement. Hyperspectral data shows different relations with LAI under various terrain conditions. LiDAR derived LAI was test as a training data-set for hyperspectral data.

 

Request a quote for hyperspectral optical components and systems

Applications of McIDAS-V to multi- and hyperspectral data analysis
Paper 7456-35 of Conference 7456
Date: Thursday, 06 August 2009
Time: 11:50 AM – 12:10 PM

Author(s): Thomas H. Achtor, Thomas D. Rink, Thomas M. Whittaker, Univ. of Wisconsin-Madison (United States)

 
The fifth generation of the Man computer Interactive Data Access System (McIDAS-V) is a java-based, open-source, freely available software package that provides powerful new capabilities to analyze and visualize data from the next generation of remote sensing instruments under development for the NPOESS and GOES-R programs. McIDAS-V provides software tools with expanded capability and performance to support innovative techniques for developing algorithms, visualizing data and products, and validating results.

This presentation will utilize MODIS and AIRS data as well as MSG SEVERI and METOP IASI data to demonstrate specific examples of analysis and product validation techniques with multi and hyper -spectral data. Case studies will demonstrate the capabilities of McIDAS-V for working on scales from high resolution severe local storms to global analysis.

Comparative analysis of different implementations of a parallel algorithm for automatic target detection and classification of hyperspectral images
Paper 7455-32 of Conference 7455
Date: Wednesday, 05 August 2009
Time: 4:20 PM – 4:40 PM

Author(s): Abel Paz, Antonio J. Plaza, Javier Plaza, Univ. de Extremadura (Spain)

 
In the last few years, several algorithms have been developed for automatic target detection in hyperspectral images, including the automatic target detection and classification (ATDCA) algorithm, which uses an orthogonal subspace projection (OSP) approach. In this paper, we develop computationally efficient parallel versions for near real-time exploitation of the ATDCA algorithm based on the use of several distance metrics in addition to the OSP approach. The parallel versions are quantitatively compared using hyperspectral data collected by NASA's AVIRIS sensor over the World Trade Center in New York, five days after the terrorist attacks.

Estimation of canopy water content with MODIS spectral indexes
Paper 7454-32 of Conference 7454
Date: Thursday, 06 August 2009
Time: 3:50 PM – 4:10 PM

Author(s): Zuoning Jiang, Lin Li, Indiana Univ.-Purdue Univ. Indianapolis (United States); Susan L. Ustin, Univ. of California, Davis (United States)

 
Equivalent water thickness (EWT) derived from AVIRIS data in the process of atmospheric correction has been proven highly correlated with in situ measured canopy water content. Four vegetation indexes, NDWI, SIWSI, NDVI, and EVI were derived from MODIS reflectance and were calibrated with hyperspectral EWT and validated with both hyperspectral EWT dataset and field measured EWT in the SMEX04 site. Additional work includes discussing factors affecting the performance of MODIS indexes and applying validated relationships between MODIS spectral indexes and vegetation water content to a time series of MODIS images to determine the temporal change of vegetation water content.

A new spaceborne hyperspectral instrument (Hyper-Spectrum Imager, HSI) aboard HJ-1-A satellite: application in retrieving atmospheric water vapor
Paper 7457-6 of Conference 7457
Date: Monday, 03 August 2009
Time: 11:30 AM – 11:50 AM

Author(s): Zifeng Wang, Liangfu Chen, Institute of Remote Sensing Applications (China); Qing Li, Ministry of Environmental Protection (China); Lin Su, Dong Han, Institute of Remote Sensing Applications (China)

 
HJ-1-A is a newly launched environmental satellite with hyperspectral instrument, that is, Hyper-Spectrum Imager (HSI). HSI is an interference imager with 115 bands located between 460nm and 950nm and its main characteristics are introduced in this paper.

To retrieve water vapor content, a prototype algorithm is developed for HSI based on the radiant transformation simulations, and it employs the combination of 3 pairs of absorption channel and window channel of HSI. The performance of the algorithm is validated by comparing HSI water vapor retrieval with MODIS water vapor product, and there is relevance between the two water vapor amounts. The HSI’s potential of retrieving water vapor is evaluated on the basis of these comparisons.

Course: Introduction to Optical Remote Sensing Systems
Date: Wednesday, 05 August 2009
Time: 1:30 PM – 5:30 PM

Instructor(s): Joseph A. Shaw, Montana State Univ./Bozeman (United States)



Remote sensing techniques to monitor nitrogen-driven carbon dynamics in vegetation
Paper 7454-2 of Conference 7454
Date: Wednesday, 05 August 2009
Time: 9:00 AM – 9:20 AM

Author(s): Lawrence A. Corp, Science Systems and Applications, Inc. (United States); Elizabeth M. Middleton, NASA Goddard Space Flight Ctr. (United States); Yen-Ben Cheng, Science Systems and Applications, Inc. (United States); Petya K. E. Campbell, Karl F. Huemmrich, NASA Goddard Space Flight Ctr. (United States); Craig S. T. Daughtry, U.S. Dept. of Agriculture (United States)


 

Development of urban surface models for improved aerosol retrieval
Paper 7456-11 of Conference 7456
Date: Wednesday, 05 August 2009
Time: 1:20 PM – 1:40 PM

Author(s): Min Oo, Matthias Jerg, Ana J. Picon, Eduardo Hernandez, Barry M. Gross, Fred Moshary, Samir A. Ahmed, The City College of New York (United States)

 
A combination of CIMEL raduiometer and MODIS measurements are used to correct surface albedo models. In particular, we show through an analysis of hyperspectral high resolution Hyperion data that the correlation coefficient assumption underestimates ground albedo resulting in an overestimate of the VIS optical depth and operational collect 5 surface model shows an incorrect trend between the MVI index and the surface correlations. Preliminary radiative transfer calculations based on the same model show that this mechanism can help explain the observed overestimation and the corrected models have been implemented for NYC and Mexico City with significantly improved AOD.

On the dispersive properties of HgCdTe as a function of depth
Paper 7419B-15 of Conference 7419B
Date: Thursday, 06 August 2009
Time: 3:40 PM – 4:00 PM

Author(s): Vaidya Nathan, Air Force Research Lab. (United States); Yong Chang, Univ. of Illinois at Chicago (United States); Paul D. LeVan, Air Force Research Lab. (United States)

 
An overview of the properties of the absorption coefficient of mercury cadmium telluride that may make this material useful for intrinsic hyperspectral detection is presented. A review of recent work in modeling the absorption coefficient is provided, and new directions for achieving higher fidelity, analytical representation are suggested.

On the dispersive properties of HgCdTe as a function of depth
Paper 7419A-15 of Conference 7419A
Date: Thursday, 06 August 2009
Time: 3:40 PM – 4:00 PM

Author(s): Vaidya Nathan, Air Force Research Lab. (United States); Yong Chang, Univ. of Illinois at Chicago (United States); Paul D. LeVan, Air Force Research Lab. (United States)

 
An overview of the properties of the absorption coefficient of mercury cadmium telluride that may make this material useful for intrinsic hyperspectral detection is presented. A review of recent work in modeling the absorption coefficient is provided, and new directions for achieving higher fidelity, analytical representation are suggested.

Request a quote for hyperspectral optical components and systems

Spectral angle mapper (SAM) for anisotropy class indexing in imaging spectrometry data
Paper 7457-10 of Conference 7457
Date: Monday, 03 August 2009
Time: 2:10 PM – 2:30 PM

Author(s): Joerg Weyermann, Univ. of Zürich (Switzerland); Daniel R. Schläpfer, RESE Applications Schläpfer (Switzerland); Andreas Hueni, Mathias Kneubuehler, Michael Schaepman, Univ. of Zürich (Switzerland)

 
The spectral angle mapper (SAM) classification algorithm is evaluated for its ability to provide anisotropy class indexing, optimized for the purpose of BRDF normalization in hyperspectral imagery. A hierarchical BRDF selection scheme is applied, focussing on the intelligent selection of BRDF-effective land cover classes. It is used with an empirical (target-) BRDF normalization method following the approach of Kennedy, published in 1997. RSL’s spectral database SPECCHIO is attached for reference spectra evaluation. Results both of the classification and the normalization process are validated using two airborne image strips from HyMAP sensor, taken over the "Vordemwald" test site in northern Switzerland.

Multicomponent compression with the latest CCSDS recommendation
Paper 7455-2 of Conference 7455
Date: Tuesday, 04 August 2009
Time: 8:40 AM – 9:00 AM

Author(s): Carole Thiebaut, Ctr. National d'Études Spatiales (France)

 
For optimum compression performances of multispectral and hyperspectral images, algorithms must exploit both spectral and spatial correlation of the data. To do this, different approaches are possible: this article proposes to use the CCSDS Image Data Compression Recommendation together with a spectral transform to perform a multicomponent compression. Depending on the number of spectral bands, an efficient spectral transform, is applied first and the CCSDS algorithm encodes each decorrelated bands. We compare the performances of such a scheme with a more optimal one including a low complexity rate or quality allocation procedure between spectrally decorrelated bands, and also with JPEG2000.

Algorithms for the categorization and identification of IR military signatures
Paper 7457-23 of Conference 7457
Date: Tuesday, 04 August 2009
Time: 11:20 AM – 11:40 AM

Author(s): Simon Turbide, Tracy L. Smithson, Daniel St-Germain, Pierre Fournier, Defence Research and Development Canada (Canada)

 
The DRDC Valcartier Spectral Imagery Laboratory currently supports two infrared ground based hyperspectral imagers (PIRATES and Baby-PIRATES) designed for field measurements of military targets. Unfortunately, military targets such as aircrafts and flares do not exhibit unique signatures. An algorithm has thus been developed to categorize target signatures based on their emission source components. A second algorithm was developed to exploit this signature description and interrogate individual field measurements for target detection and categorization. This paper highlights the major steps of each algorithm, as well as some recent results obtained for the classification and identification of flare decoy for aircraft protection.

Sparsity and morphological diversity for multivalued data analysis
Paper 7446-48 of Conference 7446
Date: Tuesday, 04 August 2009
Time: 3:40 PM – 4:00 PM

Author(s): Jerome Bobin, California Institute of Technology (United States); Jean-Luc Starck, Yassir Moudden, Commissariat à l'Énergie Atomique (France); Jalal M. Fadili, Ctr. National de la Recherche Scientifique (France)

 
Over the last few years, the development of multi-channel sensors motivated
interest in methods for the coherent processing of multivariate data. Recently, sparsity and morphological diversity
have emerged as a novel and effective source of diversity for BSS. We give here
some essential insights into the use of sparsity in source separation and we outline
the essential role of morphological diversity as being a source of diversity or contrast
between the sources. We present a new sparsity-based BSS method coined
Generalized Morphological Component Analysis (GMCA) that takes advantages of
both morphological diversity and sparsity. The proposed method has been applied to challenging
blind source separation problems ranging from astrophysical component extraction within
the ESA/Planck mission to hyperspectral data processing.

FPGA implementation of a predictor-guided lookup table method for lossless compression of 3-dimensional spectral data
Paper 7458-15 of Conference 7458
Date: Tuesday, 04 August 2009
Time: 5:40 PM – 6:00 PM

Author(s): Bormin Huang, Chia-Hsiung Chen, Allen H. Huang, Univ. of Wisconsin-Madison (United States)

 
Earlier we proposed a low-complexity algorithm, LAIS-LUT, for lossless compression of 3-D spectral data. It using lookup tables (LUTs) along with a locally averaged interband scaling (LAIS) estimate as a predictor selection mechanism. The prediction residuals are encoded with the range coder. The method applied to the 224-band AVIRIS hyperspectral image data yields a high compression ratio of 3.47 to 1.

In this paper we propose a variant of the LAIS-LUT method for more efficient implementation in Field-Programmable Gate Array (FPGA). A different predictor selection mechanism is used, and the range coder is replaced with our fast linear-time minimum-redundancy prefix coding. A preliminary FPGA prototype of the proposed 3D spectral compression chip shows significantly higher data throughput and lower power consumption than the CCSDS real-time hardware compression requirements (data throughput ≥ 20 Msamples/sec and with power consumption ≤ 0.5 watt/Msamples/sec).

FPGA implementation of a predictor-guided lookup table method for lossless compression of 3-dimensional spectral data
Paper 7455-15 of Conference 7455
Date: Tuesday, 04 August 2009
Time: 5:40 PM – 6:00 PM

Author(s): Bormin Huang, Chia-Hsiung Chen, Allen H. Huang, Univ. of Wisconsin-Madison (United States)

 
Earlier we proposed a low-complexity algorithm, LAIS-LUT, for lossless compression of 3-D spectral data. It using lookup tables (LUTs) along with a locally averaged interband scaling (LAIS) estimate as a predictor selection mechanism. The prediction residuals are encoded with the range coder. The method applied to the 224-band AVIRIS hyperspectral image data yields a high compression ratio of 3.47 to 1.

In this paper we propose a variant of the LAIS-LUT method for more efficient implementation in Field-Programmable Gate Array (FPGA). A different predictor selection mechanism is used, and the range coder is replaced with our fast linear-time minimum-redundancy prefix coding. A preliminary FPGA prototype of the proposed 3D spectral compression chip shows significantly higher data throughput and lower power consumption than the CCSDS real-time hardware compression requirements (data throughput ≥ 20 Msamples/sec and with power consumption ≤ 0.5 watt/Msamples/sec).

Assessment of Midnight Blackbody Calibration Correction (MBCC) using the Global Space-based Inter-Calibration System (GSICS)
Paper 7456-2 of Conference 7456
Date: Wednesday, 05 August 2009
Time: 8:20 AM – 8:40 AM

Author(s): Rama Varma Raja Mundakkare Kovilakom, Xiangqian Wu, National Oceanic and Atmospheric Administration (United States); Fangfang Yu, Earth Resources Technology, Inc. (United States)

 
Three-axis stabilized geostationary meteorological satellites, such as the Geostationary Operational Environmental Satellite (GOES), are susceptible to a calibration anomaly around the satellite midnight. A counter measure, the Midnight Blackbody Calibration Correction (MBCC), has been implemented. In this study, the MBCC impact is assessed with Global Space-based Inter-Calibration System (GSICS) data sets by taking advantage of the two well calibrated and highly consistent hyperspectral radiometers (Atmospheric Infrared Sounder and Infrared Atmospheric Sounding Interferometer) which are providing independent and detailed assessment of GOES calibration performance several times per day. Results of the analysis would be reported at the meeting.

GSICS GEO-LEO baseline algorithm
Paper 7456-3 of Conference 7456
Date: Wednesday, 05 August 2009
Time: 8:40 AM – 9:00 AM

Author(s): Xiangqian Wu, National Oceanic and Atmospheric Administration (United States); Tim J. Hewison, European Organisation for the Exploitation of Meteorological Satellites (Germany); Yoshihiko Tahara, Japan Meteorological Agency (Japan)

 
GSICS is a critical space component of GEOSS that provides users with high-quality inter-calibrated satellite measurements. GSICS has developed a baseline algorithm for the inter-calibration of imaging instruments on geostationary (GEO) satellites with hyperspectral sounding instruments AIRS and IASI on Low Earth Orbit (LEO) satellites. The algorithm is based on a hierarchical structure to ensure maximum consistency between all instruments subject to GSICS inter-calibration. Prototype operational algorithms have been implemented at NOAA, JMA, EUMETSAT, CMA, and KMA for inter-calibration of AIRS and IASI with GOES-11/12, MTSAT-1R, METEOSAT-7/8/9, FY-2C, and in near future the COMS. The GSICS has made significant impacts on assessing satellite data quality. This paper summarizes the major components and the theoretical basis of the algorithm.

GSICS GEO-LEO operation at NOAA/NESDIS
Paper 7456-9 of Conference 7456
Date: Wednesday, 05 August 2009
Time: 11:10 AM – 11:30 AM

Author(s): Fangfang Yu, Earth Resources Technology, Inc. (United States)

 
Global Space-based Inter-Calibration System (GSICS) is a critical space component of Global Earth Observation System of Systems (GEOSS) to provide users with high-quality inter-calibrated satellite measurements. As part of the GSICS, imaging instruments on GEO satellites have been inter-calibrated with AIRS and IASI hyperspectral sounder instruments on LEO satellites. This paper reports the GSICS GEO-LEO operation at NOAA/NESDIS, including intercalibrations of GOES-11/12 with AIRS and with IASI, and of METEOSAT-7/8/9, MTSAT-1R, and FY-2C with AIRS and IASI since August 2008. Also reviewed are algorithm development, data processing, product generation, results dissemination, and selected inter-calibration examples. The fully functioning GSICS is a powerful tool to monitor instrument performance, compare with other sensors, and diagnose calibration anomalies.

Request a quote for hyperspectral optical components and systems