A highly mobile passive remote sensing FT-IR spectrometer has been designed and built. It is based on the patented Turbo FT interferometer from D&P Instruments, and was developed for the U.S. EPA in Research Triangle Park, NC. The purpose of the work is to remotely sense, detect, and identify gaseous emissions in the field. The sensor can operate from AC power or on-board battery power. The Turbo FT is a rotary high speed Fourier Transform Infra-Red (FT-IR) spectrometer capable of operating in the rugged environments of the field. This sensor was built for 4 cm-1 resolution and 25 scans per second with a dual element MWIR/LWIR (2-16 micrometer) single pixel detector. An auto-calibrating blackbody accessory was built in, and automated real-time chemical detection software was developed. This feature allows quick calibration and facilitates the remote detection of target gas clouds. This paper will discuss the system specifications, preliminary sensor performance, and results from initial testing.

A high-speed passive FT-IR imaging spectrometer has been designed and built. This sensor, based on the patented Turbo FT interferometer from D&P Instruments, was developed under contract with the U.S. Army Edgewood Chemical and Biological Center (ECBC) in Aberdeen, MD. The Turbo FT is a rotary high speed Fourier Transform Infra-Red (FT-IR) spectrometer which has been previously run at speeds up to 360 scans per second, at up to 1 cm-1 resolution. It has also operated in rugged environments, including helicopter and fixed wing platforms. This sensor was run at 8 cm-1 resolution and 30 scans per second with an 8X8 (64 element) LWIR (8-12 micron) mosaic detector. An on-board auto-calibrating blackbody accessory was built in, and automated chemical detection software was developed. These features allow in-flight calibration, and facilitate remote detection of target gas clouds. This paper will discuss the system specifications, preliminary sensor performance, and results from initial testing.

Advance of digital electronic technology into the signal processing domain facilitates significant advancement in the ability to control the scanning mechanisms of Fourier Transform Spectrometers (FTS). A generic digital controller for FTS has been developed and is now being offered as a commercial product to upgrade most commercially available FTS that use a He-Ne laser as their metrology system. This controller replaces the conventional analog signal from the laser fringes with a digital signal using a dedicated DSP and creates a more precise feedback control of the FTS scan mechanism.

Raman spectroscopy is a well understood phenomenon and can be useful for remote material identification. Raman spectroscopy is performed by directing a laser (pump) beam onto a specimen, an extended scene, to induce Raman scatter. Since Raman scatter is a relatively weak phenomenon, a telescope is often used to collect the scattered signal and a narrow band filter is used to reject the pump scatter. The Raman scatter is processed using a spectrometer to identify the Raman signal. This spectrometer could be a dispersive (grating) spectrometer or a Fourier Transform Imaging Spectrometer (FTIS) using a traditional Michelson interferometer. We propose an experiment using an FTIS but with a Fizeau interferometer that takes the form of a multi-aperture imaging system to identify the Raman scattering. An advantage to using an FTIS with a Fizeau interferometer is it occurs naturally in a multi-aperture imaging system, i.e., no additional hardware is needed obtain spectral information. Therefore, a multi-aperture system can have both high spatial and spectral resolution. In this paper, the processing of the data for the Fizeau FTIS is similar to the standard methods but can be enhanced with non-linear restoration algorithms.

The measurement of complete polarimetric parameters for a broad spectrum of wavelengths is challenging because of the multi-dimensional nature of the data and the need to chromatically separate the light under test. As a result, current methods for spectropolarimetry and imaging polarimetry are limited because they tend to be complex and/or relatively slow. Here we experimentally demonstrate an approach to measure all four Stokes parameters using three polarization gratings and four simultaneous intensity measurements, with potential to dramatically impact the varied fields of air/space-borne remote sensing, target detection, biomedical imaging/diagnosis, and telecommunications. We have developed reactive mesogen polarization gratings using simple spin-casting and holography techniques, and used them to implement a potentially revolutionary detector capable of simultaneous measurement of full polarization information at many wavelengths with no moving or tunable elements. This polarimeter design not only enables measurements over a likely bandwidth of up to 70% of the center wavelength, it is also capable of measurements at relatively high speed (MHz or more) limited only by the choice of photo-detectors and processing power of the system. The polarization gratings themselves manifest nearly ideal behavior, including diffraction efficiencies of greater than 99%, strong polarization sensitivity of the first diffraction orders, very low incoherent scattering, and suitability for visible and infrared light. Due to its simple and compact design, simultaneous measurement process, and potential for preserving image registration, this spectropolarimeter should prove an attractive alternative to current polarization detection and imaging systems.

We propose a novel grating-based all-transmissive imaging spectrograph in which the collimating and focusing optics have the same optical structure with a linear chromatic dispersion and the minimized Seidel aberrations. The imaging spectrograph is designed as a telecentric system on both object and image side. The spectral and spatial resolutions of the imaging spectrograph can be improved. A designed and an achromatic lens type are compared for demonstrating the performance.

Multi-channel gated-intensified cameras are commonly used for capturing images at ultra-high frame rates. However, the image intensifier reduces the image resolution to such an extent that the images are often unsuitable for applications requiring high quality images, such as digital image correlation. We report on the development of a new type of non-intensified multi-channel camera system that permits recording of image sequences at ultra-high frame rates at the native resolution afforded by the imaging optics and the cameras used. This camera system is based upon the use of short duration light pulses of different wavelengths for illumination of the target and the use of wavelength selective elements in the imaging system to route each particular wavelength of light to a particular camera. A prototype of this camera system comprising four dual-frame cameras synchronized with four dual-cavity lasers producing laser pulses of four different wavelengths is described. The camera is built around a stereo microscope such that it can capture image sequences usable for 2D or 3D digital image correlation. The camera described herein is capable of capturing images at frame rates exceeding 100 MHz. The camera was used for capturing microscopic images of the chip-workpiece interface area during high speed machining. Digital image correlation was performed on the obtained images to map the shear strain rate in the primary-shear-zone during high speed machining.

Traditional approaches to hyperspectral target detection involve the application of detection algorithms to atmospherically compensated imagery. Rather than compensate the imagery, a more recent approach uses physical models to generate target sub-spaces. These radiance sub-spaces can then be used in an appropriate detection scheme to identify potential targets. The generation of these sub-spaces involves some 'a priori knowledge of data acquisition parameters, scene and atmospheric conditions, and possible calibration errors. Variation is allowed in the model since some parameters are difficult to know accurately. Each vector in the subspace is the result of a MODTRAN simulation coupled with a physical model. Generation of large target spaces can be computationally burdensome. This paper explores the use of statistical methods to describe such target spaces. The statistically modeled spaces can then be used to generate arbitrary radiance vectors to form a sub-space. Statistically modeled target sub-spaces, using limited training samples, were found to accurately resemble MODTRAN derived radiance vectors.

Advanced ground and space-based hyperspectral imager (HSI) concepts are being developed for a wide variety of scientific, civil, and military applications. Users and developers of these systems often require the specification of system performance in terms of receiver-operator characteristic (ROC) curves which plot probability-of-detection (POD) versus probability-of-false-alarm (PFA). In this paper we describe and illustrate the use of a scene-based modeling tool used to explore ROC curve parametric dependencies on target, scene, and HSI sensor design characteristics and detection algorithms in the visible/near infrared to shortwave infrared (VNIR/SWIR) spectral regime (i.e. from 0.4 to 2.5 microns). The magnitudes of the target and background spectral signatures are synthesized using MODTRAN; this accounts for pertinent solar elevation angle and albedo assumptions. Selected spectral input scenes (based on measured data) are used assuming imbedded spectral targets (selectable), where a fill-factor parameter is used to account for target dimension compared to sensor ground footprint. The HSI sensor sensitivity characteristics are imbedded via the noise-equivalent reflectivity difference (NE▵ρ) figure-of-merit which is computed spectrally based on a given sensor design configuration. Finally the POD, PFA and hence ROC parametrics are generated using a distinct candidate detection algorithm. The roles of scene clutter, illumination conditions, and sensor signal-to-noise ratio are made clear in simulation examples. In addition the impact of limited scene extent (limited scene pixel count) on the accuracy of the PFA predictions is noted and discussed.

This paper describes an effort to improve the modeling of the atmosphere in the Digital Imaging and Remote Sensing Image Generation (DIRSIG) model. The goal of this research can be divided into two main areas. The first is to improve the existing manner in which DIRSIG samples and references the atmosphere. The second is to give DIRSIG the ability to incorporate atmospheric inhomogeneities, as well as, the ability to accurately model them. DIRSIG has limitations in how it currently samples the atmosphere. From a geometric standpoint, it does not fully sample the energy which is scattered by the atmosphere towards the sensor (upwelled radiance). There are also other geometric issues which lead to inaccurate modeling results. One significant inaccuracy is the fact that DIRSIG can miscalculate the atmospheric effects resulting from modeling objects with non-zero altitudes. The plan is to correct this by completely reworking the procedure and geometry used by DIRSIG to sample the atmosphere. This research also addresses the effects of an inhomogeneous atmosphere and includes methods to model this variability in synthetic scenes. DIRSIG currently utilizes a single atmospheric look up table (LUT) that it references when creating an image. This LUT contains the information DIRSIG will need to predict the various radiance and transmission values for a homogeneous sky. There is no ability for DIRSIG to make one part of the sky optically thick, and the other clear. This will be remedied by having DIRSIG create a series of LUTs with different atmospheric properties that it can reference. With this ability DIRSIG can reference an optically different atmosphere depending on its viewing geometry, allowing a horizontally varying atmosphere.

Target detection algorithms for hyperspectral remote sensing have been studied for decades. The Least Square (LS) approach is one of the most widely used algorithms. It has been proved that the Noise Whitened Least Square (NWLS) can outperform the original version. But in order to have good results, the estimation of the noise covariance matrix is very important and still remains a great challenge. Many estimation methods have been proposed in the past, including spatial and frequency domain high-pass filter, neighborhood pixel subtraction, etc. In this paper, we further adopt the Fully Constrained Least Square (FCLS), which combine sum-to-one and non-negative constraints, with the NWLS and we also conduct a quantitative comparison with computer simulation of material spectrum from AVIRIS data base on the detection performance and the difference from the designed noise covariance matrix. We will also compare the results with real AVIRIS image scene.

Endmenber extraction has received increasing interests in hyperspectral image analysis. Two major issues are of interest. One is determination of endmembers, p, required to be generated and the other is generation of initial endmembers. Since most endmember extraction algorithms (EEAs) use randomly generated vectors as their initial endmembers to initialize their algorithms, their final generated endmembers are generally determined by these random initial endmembers. As a result, a different set of random initial endmembers may well likely produce a different final set of desired endmembers. This paper converts this disadvantage to an advantage and further resolves the above-mentioned two issues. Due to the random nature of initial endmembers, the proposed idea is to implement an EEA as a random algorithm so that a single run using a random set of initial endmembers is considered as a realization of a random algorithm. As a result, if an EEA is implemented several times with different sets of random initial endmembers, the intersection of their final generated endmembers in all runs should contain the desired endmembers. An EEA is then terminated when their produced intersections converge to the same set of desired endmembers. In this case, there is no need to determine the p. An EEA implemented in such a manner is called automatic EEA (AEEA). Two commonly used EEAs, pixel purity index (PPI) and N-finder algorithm (N-FINDR), are extended to AEEAs along with a new automatic ICA-based EEA. Experimental results demonstrate that the AEEA performs at least as well as their counterparts.

This paper presents a new approach to hyperspectral signature analysis, called Spectral Derivative Feature Coding (SDFC). It makes use of gradient changes in adjacent bands to characterize spectral variations so as to improve spectral discrimination and identification. In order to evaluate its performance, two binary coding methods, SPectral Analysis Manager (SPAM) and Spectral Feature-based Binary Coding (SFBC) are used to conduct comparative analysis. The experimental results demonstrate the proposed SDFC performs more effectively in capturing spectral characteristics.

The classic PCA (Principal Component Analysis) has been applied in hyperspectral imaging with varying success. One obstacle in its application is the potential physical interpretation of the principal components, which is questionable unless the principal component coefficients are nonnegative. In this paper, we show hyperspectral imaging applications of a recently developed methodology of nonnegative PCA, which overcomes this difficulty by constructing nonnegative principal components. We construct an approximation of a physics-derived target space, and suggest some interpretations of the resulting components.

The objective of this paper is to investigate the effects of dimensionality reduction on the statistical distribution of natural hyperspectral backgrounds. The statistical modeling is based on application of the multivariate t-elliptically contoured distribution to background regions which have been shown to exhibit "long-tail" behavior. Hyperspectral backgrounds are commonly represented with reduced dimensionality in order to minimize statistical redundancies in the spectral dimension and to satisfy data processing and storage requirements. In this investigation, we extend the statistical characterization of these backgrounds by modeling their Mahalanobis distance distributions in reduced dimensional space. The dimensionality reduction techniques applied in this paper include Principal Components Analysis (PCA) and spectral band aggregation. The knowledge gained from a better understanding of the effects of dimensionality reduction will be beneficial toward improving threshold selection for target detection applications. These investigations are done using hyperspectral data from the AVIRIS sensor and include spectrally homogeneous regions of interest obtained by visual interactive spatial segmentation.

Hyperspectral images are collected by hundreds of contiguous spectral channels and thus, the data volume to be processed is considered to be huge. With such high spectral resolution, spectral correlation among bands is expected to be very high. Band selection (BS) is one of common practices to reduce data volumes, while retaining desired information for data processing. This paper investigates issues of band selection and further develops various exploitation-based band prioritization criteria (BPC) which rank the hyperspectral bands in accordance with priorities measured by various applications such as detection, classification and endmember extraction. Three categories of BPC can be derived based on different design rationales, (1) second order statistics, (2) higher-order statistics, and (3) band correlation/dependence minimization or band correlation/dependence constraint. Unlike commonly used band selection techniques which do not specifically use the concept of band prioritization (BP) to select desired bands, this paper explores the idea of BP for band selection where an appropriate set of bands can be selected according to priority scores produced by spectral bands. As a result, the BPC provides a general guideline for band selection to meet various applications in hyperspectral data exploitation.

Kalman filter has been widely used in statistical signal processing for parameter estimation. Recently, a Kalman filter-based approach to spectral unmixing, referred to as Kalman filter-based linear unmixing (KFLU) was also developed for mixed pixel classification. However, its applicability to estimation and discrimination for hyperspectral signature characterization has not been explored where a hyperspectral signature is defined as a vector on a range of contiguous optical wavelengths of interest. This paper presents a new application of Kalman filtering in hyperspectral signature similarity and discrimination. In particular, it develops a Kalman filter-based signature estimator from which two Kalman filter-based discriminators can be derived for signature similarity and discrimination. The developed Kalman filter-based discriminators utilize a state equation to characterize a hyperspectral signature and a measurement equation to describe another hyperspectral signature, while the developed Kalman filter-based estimator makes use of state and measurement equations to describe the true signature and the observable signature respectively. The least squares error resulting from the Kalman filter-estimated hyperspectral signature is then used as the power for hyperspectral signature similarity and discrimination. Experimental results demonstrate that such Kalman filter-based discriminators are more effective than commonly used spectral similarity measures such as spectral angle mapper (SAM) or Euclidean distance.

Computed Tomographic Imaging Spectrometry (CTIS) is a technique which has been around for over a decade, providing snapshot measurements of datacubes as large as (x,y,λ) = (100,100,300). We discuss the difficulties with maximizing the resolution of a CTIS instrument and some new grating design ideas for realizing performance improvements.

The incentive for the 4D-IS concept was driven by the need to adequately resolve all four dimensions of data (2D spatial, spectral, and temporal) with a single, radiometrically calibrated sensor. Very fast changing phenomena are of interest; including missile exhaust plumes, missile intercept events, and lightning strikes, hypervelocity impacts, etc. Present sensor capabilities are limited to imaging sensors (producing spatial image), spectrometers (that produce a mean signature over an entire field of view with no spatial resolution), radiometers (producing in-band radiance over an entire FOV), or imaging spectrometers (or hyperspectral sensors, tunable filter type, pushbroom scanning, imaging Fourier Transform, Fabry-Perot, or CTHIS type) that produce a data cube containing spatial/spectral information but suffer from the fact that the cube acquisition process may take longer time than the temporal scale during which the event changes. The Computer Tomography Imaging Spectrometer (CTIS) is another sensor capable of 4D data collection. However, the inversion process for CTIS is computationally extensive and data processing time may be an issue in real-time applications. Hence, the 4D-IS concept with its ability to capture a full image cube at a single exposure and provide real time data processing offers a new and enhanced capability over present sensors. The 4D-IS uses a reformatter fiber optics to map a 2D image to a linear array that serves as an input slit to an imaging spectrometer. The paper describes three such instruments, a VNIR, a MWIR, and a dual band MW/LWIR. The paper describes the sensors' architecture, mapping, calibration procedures, and remapping the FPA plane into an image cube. Real-time remapping software is used to aid the operator in alignment of the sensor is described. Sample data are shown for rocket motor firings and other events.

A new approach for the design and fabrication of a miniaturized Hyperspectral imager is described. A unique and compact instrument has been developed by taking advantage of light propagation within bonded solid blocks of fused silica. The resulting microHSI is a VNIR hyperspectral sensor capable of operating in the 400-1000 nm wavelength range developed, patented, and built by NovaSol. The microHSI spectrograph weighs 12.4 oz from slit input to camera output. The microHSI can accommodate either custom foreoptics or C-mount input lenses to adapt to a wide range of fields-of-view (FOV). The prototype microHSI uses a telecentric F2.8 foreoptic, with 36 mm focal length, to cover a 15 degree FOV. It can resolve 960 spatial pixels, resulting in a 280 μrad IFOV for this particular foreoptics implementation. With a 1 nm/unbinned pixel dispersion, the spectrometer spectral resolution is 3.5 nm. Results of field and laboratory testing of the prototype microHSI are presented and show that the sensor consistently meets technical performance predictions. The prototype microHSI employs a holographic diffraction grating embedded within the optical blocks resulting in a 19% diffraction efficiency. Future units are anticipated to incorporate a blazed grating for improved throughput and SNR. The microHSI concept can be extended to operation over other wavelength regions. Designs are nearing completion for a SWIR version of the device, and a miniaturized LWIR microHSI sensor is currently at the conceptual design stage.

Acousto-optic tunable filters (AOTF's) feature all-electronic, agile, spectral tuning with a narrowband transmission function. They are rugged, solid-state devices with no moving parts. Under computer control, they can produce spectral/spatial image cube data that is ideal for hyper- and multi-spectral imaging applications. However, issues such as spectral leakage, non-uniform response over the image field, and out-of-band scatter need to examined and characterized in order to understand the performance of an AOTF imaging system. We present measurements of the spectral response of a visible AOTF imaging system and discuss the characterization and calibration steps taken to improve the system performance.

For multispectral imagery (MSI), spatial registration between bands is a very important part of the overall quality of the MSI product. For some remote sensing imagery, mis-registration of bands greater than about one-quarter of a pixel can be visually noticeable. Based on the successful registration processing developed for the Multispectral Thermal Imager (MTI) (known as edgereg), a derivative algorithm was developed for assessment of spatial registration accuracy of 4-band MSI. This algorithm has sub-pixel accuracy of a few hundredths of a pixel, and can be used to compute an image average mis-registration or process imagery in a block mode to create a mis-registration map. This paper presents the results of testing the algorithm on simulated and commercial imagery.

We are interested in using wide-field-of-view polarimetric imagers such as airborne POLDER (LOA, University of Lille) and AMPI (NASA) to study polarization signatures of surface targets. The airborne POLDER instrument has a field of view +/- 43° in along-track and +/- 51° cross-track. It records imagery through a rotating filter wheel with spectral filters, and in two spectral bands with linear polarization filters orientated at 0°, 60°, and 120° while flying at 70 m/s, generating images that need to be registered before spectra or the Stokes vectors can be computed. The atmospheric contributions, particularly in the short-wavelength visible bands, are anisotropic due to the scattering from molecules and aerosols, and the contrast is quite low, making automated image registration impossible. Thus, it is necessary to remove the polarized upwelling atmospheric contributions from the at-sensor radiances before the images can be registered, and target-leaving Stokes parameters can be derived. We accomplished the atmospheric correction by using the recently released vectorized version of the Second Simulation of the Satellite Signal in the Solar Spectrum (6Sv) from Eric Vermote et al. 1997. We wrote a front-end in IDL to run 6Sv over a range of viewing zenith and azimuth angles. The resulting Stokes parameters are then interpolated to a grid of input viewing coordinates for the sensor. Next, the Stokes hemispherical path radiances are converted to match the data of the airborne POLDER instrument for the linear polarization angles of 0°, 60°, and 120°. The upwelling atmospheric intensities are subtracted from the respective polarimetric intensity images and the difference is divided by the transmission multiplied with the solar irradiance. This atmospheric correction significantly reduces the low-frequency variations in intensity in the images resulting from atmospheric scattering. The atmospherically corrected intensity images at 0°, 60°, and 120° are then used to calculate the Stokes parameter images in the usual fashion. From the Stokes parameter images we calculate the degree and azimuth of linear polarization images for the surface.

In this paper, we investigate the use of nonlinear structure to derive the physical characteristics of coastal data. In particular, we show how the physics of shallow water coastal regions lead to well defined nonlinear structures (manifolds) in the corresponding hyperspectral data. The exact form of this structure is determined by both the Inherent Optical Properties of the water column as well as the boundary conditions (bottom reflectance, depth). This structure is then used to develop efficient algorithms for searching large 'lookup tables' of precalculated spectra with known physical characteristics, which are used for estimating the various physical parameters (bathymetry, bottom type, etc.) of the scene. We assess our methods with data collected by the NRL PHILLS sensor at the Indian River Lagoon (IRL) in Florida. The IRL is a well-studied and characterized body of water that contains a number of different water and bottom types at various shallow (generally less than 8 meters, except in the shipping channel where depths can be as much as 18 m) depths. We show in particular that the search algorithm is able to produce valid results in a short amount of time, and compare our results with an IRL LIDAR bathymetry survey from early 2004.

A previous study adapted a variety of techniques derived from multi-spectral image classification to find objects amid cluttered backgrounds in hyperspectral imagery. That study quantitatively compared the adapted algorithms against a standard object search, the matched filter (MF) and a recently developed object detector, Adaptive Cosine Estimator (ACE) and found substantial reduction in false alarm rates for a given target detection probability. One adapted object search, Regularized Maximum Likelihood Classifier (RMLC), requires calculating the covariance matrix involving the average object spectral signature and the target pixels. The object covariance matrix requires a relatively large number of pixels to generate a non-singular, accurate covariance matrix. This study examines the robustness of the RMLC algorithm on number of training pixels, the optimal mixing covariance matrices, and choice of object subspaces for the ACE algorithm. The tests were applied to visible/near IR data collected from forest and desert environments. This study finds that high detection performance standards for RMLC are invariant for pixel number for homogenous targets, down to two pixels. Regularization is relatively unaffected by the choice of areas to optimize the object covariance matrix although targets that mix background appear to be more sensitive to choice of covariance matrix. Reducing the object subspace dimensions by using the average target signature or choosing the first principle component enhances ACE performance relative to using the entire object space.

Spectral signatures derived from a multispectral or hyperspectral imager can be used in matched filter algorithms to help distinguish targets from background. In this paper we demonstrate the use of these matched filters for different target implantation models. We show that even though a specific matched filter is designated for a particular implantation model, we can use other matched filters and obtain higher detection values for low false alarm rates. We evaluate the efficiency of the algorithms by systematically implanting the target's signature into every pixel in the image and obtaining its score; the lowest scores are those pixels in which the target may be missed. For every algorithm, we generate histograms for the no-target and target cases and then analyze using the classical ROC curve.

Remote sensing of chemical warfare agents (CWA) with stand-off hyperspectral sensors has a wide range of civilian and military applications. These sensors exploit the spectral changes in the ambient photon flux produced thermal emission or absorption after passage through a region containing the CWA cloud. In this work we focus on (a) staring single-pixel sensors that sample their field of view at regular intervals of time to produce a time series of spectra and (b) scanning single or multiple pixel sensors that sample their FOV as they scan. The main objective of signal processing algorithms is to determine if and when a CWA enters the FOV of the sensor. We shall first develop and evaluate algorithms for staring sensors following two different approaches. First, we will assume that no threat information is available and we design an adaptive anomaly detection algorithm to detect a statistically-significant change in the observed spectrum. The algorithm processes the observed spectra sequentially-in-time, estimates adaptively the background, and checks whether the next spectrum differs significantly from the background based on the Mahalanobis distance or the distance from the background subspace. In the second approach, we will assume that we know the spectral signature of the CWA and develop sequential-in-time adaptive matched filter detectors. In both cases, we assume that the sensor starts its operation before the release of the CWA; otherwise, staring at a nearby CWA-free area is required for background estimation. Experimental evaluation and comparison of the proposed algorithms is accomplished using data from a long-wave infrared (LWIR) Fourier transform spectrometer.

Anomaly detection for remote sensing has been intensely investigated in recent years. It is not an easy task since an anomaly has distinct unknown spectral features from its neighborhood, and it usually has small size with only a few pixels. Several methods are devoted to this problem, such as the well-known RX algorithm which takes advantage of the second-order statistics. The RX algorithm assumes Gaussian noise and uses sample covariance matrix for data whitening. However, when the anomalies pixel number exceeds certain percentage or the data is ill distributed, the sample covariance matrix can not represent the background distribution. In this case, the RX algorithm will not perform well. In this paper, we perform a computer simulation to analyze the performance of the RX algorithm under different circumstances, including the number of anomaly pixels, number of anomaly types, the distance of anomaly spectrum from background, the noise distribution, etc. Later we used AVIRIS data and utilized the characteristic of principle component analysis to estimate the covariance matrix and mean of the pixels of the background. We will analyze the performance of the RX algorithm by using the estimated covariance matrix with the original version.

In preparation for future atmospheric space missions a consortium of Dutch organizations is performing design studies on a nadir viewing grating-based imaging spectrometer using OMI and SCIAMACHY heritage. The spectrometer measures selected species (O₃, NO₂, HCHO, H₂O, SO₂, aerosols (optical depth, type and absorption index), CO and CH4) with sensitivity down to the Earth's surface, thus addressing science issues on air quality and climate. It includes 3 UV-VIS channels continuously covering the 270-490 nm range, a NIR-channel covering the 710-775 nm range, and a SWIR-channel covering the 2305-2385 nm range. This instrument concept is, named TROPOMI, part of the TRAQ-mission proposal to ESA in response to the Call for Earth Explorer Ideas 2005, and, named TROPI, part of the CAMEO-proposal prepared for the US NRC decadal study-call on Earth science and applications from space. The SWIR-channel is optional in the TROPOMI/TRAQ instrument and included as baseline in the TROPI/CAMEO instrument. This paper focuses on derivation of the instrument requirements of the SWIR-channel by presenting the results of retrieval studies. Synthetic detector spectra are generated by the combination of a forward model and an instrument simulator that includes the properties of state-of-the-art detector technology. The synthetic spectra are input to the CO and CH₄ IMLM retrieval algorithm originally developed for SCIAMACHY. The required accuracy of the Level-2 SWIR data products defines the main instrument parameters like spectral resolution and sampling, telescope aperture, detector temperature, and optical bench temperature. The impact of selected calibration and retrieval errors on the Level-2 products has been characterized. The current status of the SWIR-channel optical design with its demanding requirements on ground-pixel size, spectral resolution, and signal-to-noise ratio will be presented.

For exploration of planetary surfaces, detection of water and ice is of great interest in supporting existence of life on other planets. Therefore, a remote Raman spectroscopy system was demonstrated at NASA Langley Research Center in collaboration with the University of Hawaii for detecting ice-water and hydrous minerals on planetary surfaces. In this study, a 532 nm pulsed laser is utilized as an excitation source to allow detection in high background radiation conditions. The Raman scattered signal is collected by a 4-inch telescope positioned in front of a spectrograph. The Raman spectrum is analyzed using a spectrograph equipped with a holographic super notch filter to eliminate Rayleigh scattering, and a holographic transmission grating that simultaneously disperses two spectral tracks onto the detector for higher spectral range. To view the spectrum, the spectrograph is coupled to an intensified charge-coupled device (ICCD), which allows detection of very weak Stokes line. The ICCD is operated in gated mode to further suppress effects from background radiation and long-lived fluorescence. The sample is placed at 5.6 m from the telescope, and the laser is mounted on the telescope in a coaxial geometry to achieve maximum performance. The system was calibrated using the spectral lines of a Neon lamp source. To evaluate the system, Raman standard samples such as calcite, naphthalene, acetone, and isopropyl alcohol were analyzed. The Raman evaluation technique was used to analyze water, ice and other hydrous minerals and results from these species are presented.