This PDF file contains the front matter associated with SPIE Proceedings Volume 9611, including the Title Page, Copyright information, Table of Contents, Introduction (if any), and Conference Committee listing.

The Snow and Water Imaging Spectrometer (SWIS) is a fast, high-uniformity, low-polarization sensitivity imaging spectrometer and telescope system designed for integration on a 6U CubeSat platform. Operating in the 350-1700 nm spectral region with 5.7 nm sampling, SWIS is capable of simultaneously addressing the demanding needs of coastal ocean science and snow/ice monitoring. We discuss progress in the SWIS optomechanical design, thermal analysis, and mission plan. We also describe an innovative single drive on-board calibration system capable of addressing the stringent radiometric stability and knowledge these missions require. The spectrometer features a new Teledyne CHROMA array, optimized for high temperature operation, with a linear variable anti-reflection coating to enhance quantum efficiency and minimize backscatter.

Mud samples collected from bodies of water reveal information about the distribution of microorganisms in the local sediments. Hyperspectral imaging has been investigated as a technology to identify phototropic organisms living on sediments collected from the Texas Coastal Bend area based on their spectral pigment profiles and spatial arrangement. The top pigment profiles identified through high-performance liquid chromatography (HPLC) have been correlated with spectral signatures extracted from the hyperspectral data of mud using fast Fourier transform (FFT). Spatial distributions have also been investigated using 2D hyperspectral image processing. 2D pigment distribution maps have been created based on the correlation with pigment profiles in the FFT domain. Among the tested pigments, the results show match among four out of five pigment distribution trends between HPLC and hyperspectral data analysis. Differences are attributed mainly to the difference between area and volume of scale between the HPLC analysis and area covered by hyperspectral imaging.

In recent years long-wavelength infrared (LWIR) hyperspectral imagery has significantly improved in quality and become much more widely available, sparking interest in a variety of applications involving remote sensing of surface composition. This in turn has motivated the development and study of LWIR-focused algorithms for atmospheric retrieval, temperature-emissivity separation (TES) and material detection and identification. In this paper we evaluate some LWIR algorithms for atmospheric retrieval, TES, endmember-finding and rare material detection for their utility in characterizing mineral composition in SEBASS hyperspectral imagery taken near Cuprite, NV. Atmospheric correction results using the In-Scene Atmospheric Correction (ISAC) method are compared with those from the first-principles, MODTRAN©-based FLAASH-IR method. Covariance-whitened endmember-finding methods are observed to be sensitive to image artifacts. However, with clean data and all-natural terrain they can automatically locate and distinguish many minor mineral components, with especially high sensitivity to varieties of calcite. Not surprisingly, the major scene materials, including silicates, are best located using unwhitened techniques. Minerals that we identified in the data include calcite, quartz, alunite and (tentatively) kaolinite.

The Atmospheric Infrared Sounder (AIRS) is a hyperspectral infrared instrument on the EOS Aqua Spacecraft, launched on May 4, 2002. AIRS has 2378 infrared channels ranging from 3.7 μm to 15.4 μm and a 13.5 km footprint at nadir. AIRS is a “facility” instrument developed by NASA as an experimental demonstration of advanced technology for remote sensing and the benefits of high resolution infrared spectra to science investigations. AIRS, in conjunction with the Advanced Microwave Sounding Unit (AMSU), produces temperature profiles with 1K/km accuracy on a global scale, as well as water vapor profiles and trace gas amounts for CO₂ , CO, SO₂ , O₃ and CH₄. AIRS data are used for weather forecasting, climate process studies and validating climate models. The AIRS instrument has far exceeded its required design life of 5 years, with nearly 13 years of routine science operations that began on August 31, 2002. While the instrument has performed exceptionally well, with little sign of wear, the AIRS Project continues to monitor and maintain the health of AIRS, characterize its behavior and improve performance where possible. Radiometric stability has been monitored and trending shows better than 16 mK/year stability. Spectral calibration stability is better than 1 ppm/year. At this time we expect the AIRS to continue to perform well into the next decade. This paper contains updates to previous instrument status reports, with emphasis on the last three years.

Food safety and quality in packaged products are paramount in the food processing industry. To ensure that packaged products are free of foreign materials, such as debris and pests, unwanted materials mixed with the targeted products must be detected before packaging. A portable hyperspectral imaging system in the visible-to-NIR range has been used to acquire hyperspectral data cubes from pinto beans that have been mixed with foreign matter. Bands and band ratios have been identified as effective features to develop a classification scheme for detection of foreign materials in pinto beans. A support vector machine has been implemented with a quadratic kernel to separate pinto beans and background (Class 1) from all other materials (Class 2) in each scene. After creating a binary classification map for the scene, further analysis of these binary images allows separation of false positives from true positives for proper removal action during packaging.

This research uses General Circulation Model (GCM) derived products, with 1 km spatial resolution and sampled every 10 minutes, over a moving area following the track of a simulated severe Atlantic storm. Model products were aggregated over sounder footprints corresponding to 13 km in LEO, 2 km in LEO, and 5 km in GEO sampled every 72 minutes. We simulated radiances for instruments with AIRS-like spectral coverage, spectral resolution, and channel noise, using these aggregated products as the truth, and analyzed them using a slightly modified version of the operational AIRS Version-6 retrieval algorithm. Accuracy of retrievals obtained using simulated AIRS radiances with a 13 km footprint was similar to that obtained using real AIRS data. Spatial coverage and accuracy of retrievals are shown for all three sounding scenarios. The research demonstrates the potential significance of flying Advanced AIRS-like instruments on future LEO and GEO missions.

The performance analysis of a satellite mission requires specific tools that can simulate the behavior of the platform; its payload; and the acquisition of scientific data from synthetic scenes. These software tools, called End-to-End Mission Performance Simulators (E2ES), are promoted by the European Space Agency (ESA) with the goal of consolidating the instrument and mission requirements as well as optimizing the implemented data processing algorithms. Nevertheless, most developed E2ES are designed for a specific satellite mission and can hardly be adapted to other satellite missions. In the frame of ESA's FLEX mission activities, an E2ES is being developed based on a generic architecture for passive optical missions. FLEX E2ES implements a state-of-the-art synthetic scene generator that is coupled with dedicated algorithms that model the platform and instrument characteristics. This work will describe the flexibility of the FLEX E2ES to simulate complex synthetic scenes with a variety of land cover classes, topography and cloud cover that are observed separately by each instrument (FLORIS, OLCI and SLSTR). The implemented algorithms allows modelling the sensor behavior, i.e. the spectral/spatial resampling of the input scene; the geometry of acquisition; the sensor noises and non-uniformity effects (e.g. stray-light, spectral smile and radiometric noise); and the full retrieval scheme up to Level-2 products. It is expected that the design methodology implemented in FLEX E2ES can be used as baseline for other imaging spectrometer missions and will be further expanded towards a generic E2ES software tool.

This paper provides best practices for the radiometric performance modeling of imaging spectrometers. A set of standard terminology is proposed to use when modeling imaging spectrometers. The calculation of various radiometric sensitivity metrics and their contrast counterparts are outlined. Modeling approaches are described for both solar reflected and thermally emitted bands. Finally, this approach is applied to an example hyperspectral sensor.

The microfacet BRDF model is preferred to describe reflectance in many applications due to its closed-form approximation to the BRDF which is relatively easy to use; however, it almost entirely excludes wavelength-dependent scaling of the reflectance distribution. To rectify this, the BRDF was measured at multiple incident angles and for multiple materials at several wavelengths between 3.39 μm and 10.6 μm. Results quantify the dramatic change in the specular BRDF of a variety of materials even after accounting for overall reflectance, and suggests it is necessary to modify the wavelength dependence in the microfacet model.

The sensitivity of hyper-spectral remote sensing to the directional reflectance of surfaces was studied using both laboratory and field measurements. Namely, the effects of the specular- and diffuse-reflectance properties of a set of eight samples, ranging from high to low in both total reflectance and specularity, on diffuse-only and diffusespecular radiative transfer models in the long-wave infrared (LWIR, 7-14-μm wavelength) were studied. The samples were measured in the field as a set of eight panels, each in two orientations, with surface normal pointing toward zenith and tipped at 45° from zenith. The field-collected data also included down-welling spectral sky radiance at several angles from zenith to the horizon, ground spectral radiance, panel spectral radiances in both orientations, Infragold® spectral radiances in both orientations near each panel location, and panel temperatures. Laboratory measurements included spectral hemispherical, specular and diffuse directional reflectance (HDR, SDR and DDR) for each sample for several reflectance angles with respect to the surface normal. The diffuse-only radiative transfer model used the HDR data, while the diffuse-specular model used the SDR and DDR data. Both calculated spectral reflected and self-emitted radiances for each panel, using the field-collected sky radiance data to avoid uncertainties associated with atmospheric models. The modeled spectral radiances were then compared to the field-collected values to quantify differences in moving from an HDR-based model to an SDR/DDR model in the LWIR for a variety of surface-reflectance types.

Approximate solutions to the Radiative transfer equation for granular media have been previously developed¹. To apply these models to coastal sediments, modifications are needed to account for observed phenomenology. This study uses a new hyperspectral goniometer system, the Goniometer of the Rochester Institute of Technology (GRIT), designed for both field and laboratory settings, to compare observed bidirectional reflectance distribution function (BRDF) measurements with outcomes predicted by the approximate radiative transfer solutions. In previous laboratory studies,2 using a more limited hyperspectral goniometer observing in the principle plane, we had seen that the degree of optical contrast between coastal sand constituents was indicative of whether these models accurately predict the observed BRDF dependence on sediment density. Our earlier measurements using another field hyperspectral goniometer also demonstrated results consistent with the laboratory measurements as well as with CASI- 1500 airborne hyperspectral measurements^3,4. In our earlier work,² the presence of highly contrasting constituents (translucent quartz and more opaque fractions composed of minerals such as magnetite) led to greater reflectance as density decreased, exactly the opposite of what was anticipated from radiative transfer models for a more uniform sand. The present study shows that the illumination zenith angle also plays a significant role in whether or not BRDF dependency exhibits behavior predicted by current radiative transfer theory, and this distinction is directly related to the degree of multiple scattering, which depends on the illumination zenith angle. We also investigate a novel sampling paradigm that constrains the measurements to constant phase angle and reveals when the multiple scattering component of models departs from the assumptions of current theory. For the multiple scattering term, we also propose and analyze a simple modification which removes the isotropic assumption and provides a better match to BRDF observations under the constrained sampling paradigm.

In this paper we present a comparative analysis of three vision systems to nondestructively predict defects on the surfaces of aluminum castings. A hyperspectral imaging system, a thermal imager, and a digital color camera have been used to inspect aluminum metal cast surfaces. Hyperspectral imaging provides both spectral and spatial information, as each material produces specific spectral signatures which are also affected by surface texture. Thermal imager detects infrared radiation whereby hotspots can be investigated to identify possible trapped inclusions close to the surface, or other superficial defects. Finally, digital color images show apparent surface defects that can also be viewed with the naked eye but can be automated for fast and efficient data analysis. The surface defect locations predicted using the three systems are then verified by breaking the casings using a tensile tester. Of the three nondestructive methods, the thermal imaging camera was found to produce the most accurate predictions for defect location that caused breakage.

The spectral reflectance of a sample of quartz sand was monitored as the sample progressed from air-dry to fully saturated, and then back to air-dry. Wetting was accomplished by spraying small amounts of water on the surface of the sample, and collecting spectra whenever change occurred. Drying was passive, driven by evaporation from the sand surface, with spectra collected every 5 minutes until the sample was air dry. Water content was determined by monitoring the weight of the sample through both wetting and drying. There was a pronounced difference in the pattern of change in reflectance during wetting and drying, with the differences being apparent both in spectral details (i.e., the depth of absorption bands) and in the magnitude of the reflectance for a particular water content. The differences are attributable to the disposition of water in the sample. During wetting, water initially occurred only on the surface, primarily as water adsorbed onto sand particles. With increased wetting the water infiltrated deeper into the sample, gradually covering all particles and filling the pore spaces. During drying, water and air were distributed throughout the sample for most of the drying period. The differences in water distribution are assumed to be the cause of the differences in reflectance and to the differences in the depths of four strong water absorption bands.

Spontaneous Raman scattering is an effective technique in gas analysis, but the detection of minor constituents is difficult because of the low signal level and the usually existed background. Imaging spectrometer can provide highly spatial resolved spectra, so it should be much easier to pick up Raman signal of minor constituents from the Raman/fluorescence background of the sample cell and transporting optics compared with the widely used fiber-coupled spectrometers. For this reason, an imaging spectrometer was constructed from transmitting volume phase holographic grating, camera lenses and CCD detector. When it was used to analyze the gas sample in metal-lined capillary, which is a sample cell believed with great enhancement of Raman signal, the background was compressed obviously. When it was used to analyze the gas in a sample cell including a parabolic reflector, only weak background signal was observed, as the wide separation between the collecting zone (the focus point of the parabolic surface) and the wall of sample cell benefitted to the analysis by imaging spectrometer. By using the last sample cell, the signal from CO₂ in ambient air was able to be found by an exposure time about 20 sec, and limits of detection for H₂, CO₂ and CO were estimated as 60 ppm, 100 ppm and 300 ppm respectively by the results of a longer exposure time. These results show that an imaging spectrometer paired with a well-arranged sample cell will lower the detecting limit effectively.

We show the miniaturization and parallelization of a scanning standing wave spectrometer with a long term goal of creating a compact imaging spectrometer. In our standing wave integrated Fourier transform spectrometer, light is injected with micro-lenses into several optical polymer waveguides. A piezo actuated mirror located at the waveguide end-facet can shift the interferogram to increase its sampling frequency. The spatial distribution of the standing wave intensity inside the waveguide is partially scattered out of the plane by a periodic metallic grating and recorded by a CCD camera. We present spectra acquisition for six adjacent waveguides simultaneously at a wavelength of 632.8 nm.

The proliferation of lasers has led to their widespread use in applications ranging from short range standoff chemical detection to long range Lidar sensing and target designation operating across the UV to LWIR spectrum. Recent advances in high energy lasers have renewed the development of laser weapons systems. The ability to measure and assess laser source information is important to both identify a potential threat as well as determine safety and nominal hazard zone (NHZ). Laser detection sensors are required that provide high dynamic range, wide spectral coverage, pulsed and continuous wave detection, and large field of view. OPTRA, Inc. and Tufts have developed a custom ROIC smart pixel imaging sensor architecture and wavelength encoding optics for measurement of source wavelength, pulse length, pulse repetition frequency (PRF), irradiance, and angle of arrival. The smart architecture provides dual linear and logarithmic operating modes to provide 8+ orders of signal dynamic range and nanosecond pulse measurement capability that can be hybridized with the appropriate detector array to provide UV through LWIR laser sensing. Recent advances in sputtering techniques provide the capability for post-processing CMOS dies from the foundry and patterning PbS and PbSe photoconductors directly on the chip to create a single monolithic sensor array architecture for measuring sources operating from 0.26 – 5.0 microns, 1 mW/cm² – 2 kW/cm².

The uncertainties in the knowledge of the Instrument Spectral Response Function (ISRF), barycenter of the spectral channels and bandwidth / spectral sampling (spectral resolution) are important error sources in the processing of satellite imaging spectrometers within narrow atmospheric absorption bands. The exhaustive laboratory spectral characterization is a costly engineering process that differs from the instrument configuration in-flight given the harsh space environment and harmful launching phase. The retrieval schemes at Level-2 commonly assume a Gaussian ISRF, leading to uncorrected spectral stray-light effects and wrong characterization and correction of the spectral shift and smile. These effects produce inaccurate atmospherically corrected data and are propagated to the final Level-2 mission products. Within ESA's FLEX satellite mission activities, the impact of the ISRF knowledge error and spectral calibration at Level-1 products and its propagation to Level-2 retrieved chlorophyll fluorescence has been analyzed. A spectral recalibration scheme has been implemented at Level-2 reducing the errors in Level-1 products below the 10% error in retrieved fluorescence within the oxygen absorption bands enhancing the quality of the retrieved products. The work presented here shows how the minimization of the spectral calibration errors requires an effort both for the laboratory characterization and for the implementation of specific algorithms at Level-2.

Spectral imaging systems lead to enhanced sensing properties when the sensing system provides sufficient spectral resolution to identify materials from its spectral reflectance signature. The performance of diffraction gratings provides an initial way to improve instrumental resolution. Thus, subsequent manufacturing techniques of high quality gratings are essential to significantly improve the spectral performance. The ZEISS unique technology of manufacturing real-blazed profiles comprising transparent substrates is well suited for the production of transmission gratings. In order to reduce high order aberrations, aspherical and free-form surfaces can be alternatively processed to allow more degrees of freedom in the optical design of spectroscopic instruments with less optical elements and therefore size and weight advantages. Prism substrates were used to manufacture monolithic GRISM elements for UV to IR spectral range. Many years of expertise in the research and development of optical coatings enable high transmission anti-reflection coatings from the DUV to the NIR. ZEISS has developed specially adapted coating processes (Ion beam sputtering, ion-assisted deposition and so on) for maintaining the micro-structure of blazed gratings in particular. Besides of transmission gratings, numerous spectrometer setups (e.g. Offner, Rowland circle, Czerny-Turner system layout) working on the optical design principles of reflection gratings. This technology steps can be applied to manufacture high quality reflection gratings from the EUV to the IR applications with an outstanding level of low stray light and ghost diffraction order by employing a combination of holography and reactive ion beam etching together with the in-house coating capabilities. We report on results of transmission, reflection gratings on plane and curved substrates and GRISM elements with enhanced efficiency of the grating itself combined with low scattered light in the angular distribution. Focusing on the straylight characteristic a measurement of the actual straylight level, preferably with extremely high precision, was performed and will be discussed in this paper. Beside of the results of straylight measurement the actual results on improving efficiency for transmission and reflection gratings will be discussed on theoretical simulations compared to measured data over the entire wavelength range.

Exploitation of longwave infrared hyperspectral imagery often requires atmospheric compensation in order to retrieve intrinsic material emissivity properties. For materials possessing subtle spectral features and those with lower emissivity, downwelling radiance plays an important role in the atmospheric compensation process. Most atmospheric compensation algorithms use an estimate of the total downwelling radiance integrated over the hemisphere. However, for tilted surfaces and non-nadir imaging scenarios, directional downwelling radiance information may be required. This work examines collection of directional downwelling radiance measurements using a Fourier transform infrared spectrometer. Specifically, work is done to determine a minimum number of sky angles to measure for which the remainder of the directional sky measurements can be estimated with minimal error through the use of simple data fitting models.

Hyperspectral imaging for remote sensing has prompted development of hyperspectral image projectors that can be used to characterize hyperspectral imaging cameras and techniques in the lab. One such emerging astronomical hyperspectral imaging technique is wide-field double-Fourier interferometry. NASA’s current, state-of-the-art, Wide-field Imaging Interferometry Testbed (WIIT) uses a Calibrated Hyperspectral Image Projector (CHIP) to generate test scenes and provide a more complete understanding of wide-field double-Fourier interferometry. Given enough time, the CHIP is capable of projecting scenes with astronomically realistic spatial and spectral complexity. However, this would require a very lengthy data collection process. For accurate but time-efficient projection of complicated hyperspectral images with the CHIP, the field must be decomposed both spectrally and spatially in a way that provides a favorable trade-off between accurately projecting the hyperspectral image and the time required for data collection. We apply nonnegative matrix factorization (NMF) to decompose hyperspectral astronomical datacubes into eigenspectra and eigenimages that allow time-efficient projection with the CHIP. Included is a brief analysis of NMF parameters that affect accuracy, including the number of eigenspectra and eigenimages used to approximate the hyperspectral image to be projected. For the chosen field, the normalized mean squared synthesis error is under 0.01 with just 8 eigenspectra. NMF of hyperspectral astronomical fields better utilizes the CHIP’s capabilities, providing time-efficient and accurate representations of astronomical scenes to be imaged with the WIIT.

There are a multitude of civilian and military applications for the detection of anomalous changes in hyper-spectral images. Anomalous changes occur when the material within a pixel is replaced. Environmental factors that change over time, such as illumination, will affect the radiance of all the pixels in a scene, despite the materials within remaining constant. The goal of an anomalous change detection algorithm is to suppress changes caused by the environment, and detect pixels where the materials within have changed.

Anomalous change detection is a two step process. Two co-registered images of a scene are first transformed to maximize the overall correlation between the images, then an anomalous change detector (ACD) is applied to the transformed images. The transforms maximize the correlation between the two images to attenuate the environmental differences that distract from the anomalous changes of importance.

Several categories of transforms with different optimization parameters are discussed and compared. One of two types of ACDs are then applied to the transformed images. The first ACD uses the difference of the two transformed images. The second concatenates the spectra of two images and uses an aggregated ACD. A comparison of the two ACD methods and their effectiveness with the different transforms is done for the first time.

We present image processing algorithms that improve spatial and spectral resolution on the Snapshot Hyperspectral Imaging Fourier Transform (SHIFT) spectrometer. Final measurements are stored in the form of threedimensional datacubes containing the scene’s spatial and spectral information. We discuss calibration procedures, review post-processing methods, and present preliminary results from proof-of-concept experiments.

Linear mixtures of materials in a scene often occur because the pixel size of a sensor is relatively large and consequently they contain patches of different materials within them. This type of mixing can be thought of as areal mixing and modeled by a linear mixture model with certain constraints on the abundances. The solution to these models has received a lot of attention. However, there are more complex situations, such as scattering that occurs in mixtures of vegetation and soil, or intimate mixing of granular materials like soils. Such multiple scattering and microscopic mixtures within pixels have varying degrees of non-linearity. In such cases, a linear model is not sufficient. Furthermore, often enough, scenes may contain cases of both linear and non-linear mixing on a pixel-by-pixel basis. This study considers two approaches for use as generalized methods for un-mixing pixels in a scene that may be linear (areal mixed) or non-linear (intimately mixed). The first method is based on earlier studies that indicate non-linear mixtures in reflectance space are approximately linear in albedo space. The method converts reflectance to singlescattering albedo (SSA) according to Hapke theory assuming bidirectional scattering at nadir look angles and uses a constrained linear model on the computed albedo values. The second method is motivated by the same idea, but uses a kernel that seeks to capture the linear behavior of albedo in non-linear mixtures of materials. The behavior of the kernel method is dependent on the value of a parameter, gamma. Furthermore, both methods are dependent on the choice of endmembers, and also on RMSE (root mean square error) as a performance metric. This study compares the two approaches and pays particular attention to these dependencies. Both laboratory and aerial collections of hyperspectral imagery are used to validate the methods.

Within urban environments and other complex illumination conditions, tilted target surface or partial sky occlusions from nearby raised objects can have a significant impact on a target’s radiance observed from a spectral sensor. At the pixel level, these terms can be predicted and corrected for by modeling the impact collocated height data has on the scene radiometry. After properly accounting for these impacts, a Lambertian material’s retrieved spectral reflectivity should be the same at any location or orientation within the scene. This paper proposes a novel approach for using this constraint to iterate on atmospheric aerosol parameters until the difference of retrieved spectral reflectance of two pixels of the same material, but under different illumination conditions, is minimized.

In hyperspectral image processing, anomaly detection is a valuable way of searching targets whose spectral characteristics are not known, and the estimation of background signals is the key procedure. On account of the high dimensionality and complexity of hyperspectral image, dimensionality reduction and background suppression is necessary. In addition, the complementarity of different anomaly detection algorithms can be utilized to improve the effectiveness of anomaly detection. In this paper, we propose a novel method of anomaly detection, which is based on clustering of optimized K-means and decision-level fusion. In our proposed method, pixels with similar features are firstly clustered using an optimized k-means method. Secondly, dimensionality reduction is conducted using principle component analysis to reduce the amount of calculation. Then, to increase the accuracy of detection and decrease the false-alarm ratio, both Reed-Xiaoli (RX) and Kernel RX algorithm are used on processed image. Lastly, a decision-level fusion is processed on the detection results. A simulated hyperspectral image and a real hyperspectral one are both used to evaluate the performance of our proposed method. Visual analysis and quantative analysis of receiver operating characteristic (ROC) curves show that our algorithm can achieve better performance when compared with other classic approaches and state-of-the-art approaches.

Current algorithms of endmember extraction generally need to determine the number of endmembers manually. However, the number of endmembers is unknown in practical application, so an automated and iterative endmember extraction algorithm is put forward in this paper to solve the problem. Firstly, due to the spectral information of endmember is similar with its neighbors but noise is independent with others, we analyze the relevance between pixels and endmember in the concentric sliding window centered at each test endmember in order to eliminate the influence of noise. Then, due to the independence among endmembers, a candidate set formed of endmembers which have been extracted is constructed. We compute the correlation between the new endmember and the candidates in the set each time, if the largest correlation is small; the new one is added to the set. If the new one fails to join the set directly, we can take it to replace the existed in the set to increase the distance among endmembers. Finally, if the endmembers in the set remain unchanged in a few times, the iteration stops. The experiment shows that the improved algorithm have a near accuracy of endmember extraction with the traditional algorithm, meanwhile it weakens the influence of noise on the endmember extraction.

Dimensionality reduction is a frequent preprocessing step in hyperspectral image analysis. High-dimensional data will cause the issue of the “curse of dimensionality” in the applications of hyperspectral imagery. In this paper, a dimensionality reduction method of hyperspectral images based on random projection (RP) for target detection was investigated. In the application areas of hyperspectral imagery, e.g. target detection, the high dimensionality of the hyperspectral data would lead to burdensome computations. Random projection is attractive in this area because it is data independent and computationally more efficient than other widely-used hyperspectral dimensionality-reduction methods, such as Principal Component Analysis (PCA) or the maximum-noise-fraction (MNF) transform. In RP, the original highdimensional data is projected onto a low dimensional subspace using a random matrix, which is very simple. Theoretical and experimental results indicated that random projections preserved the structure of the original high-dimensional data quite well without introducing significant distortion. In the experiments, Constrained Energy Minimization (CEM) was adopted as the target detector and a RP-based CEM method for hyperspectral target detection was implemented to reveal that random projections might be a good alternative as a dimensionality reduction tool of hyperspectral images to yield improved target detection with higher detection accuracy and lower computation time than other methods.

Standoff detection, identification and quantification of chemicals require sensitive spectrometers with calibration capabilities. We have developed a compact novel instrument that can not only provide imaging capability, bust also one that provides spectral capability of the field of view (FOV) center under the image-guided. The system employs a Fourier transform infrared (FT-IR) spectrometer, coupled with chalcogenide glass optical fiber, and a specially designed infrared optic lens. A special kit provided by Bruker Optics is connected on the spectrometer to focus the infrared beam from the lens at the entry of the fiber. Its spectral range covers the infrared band from 1850cm^-1 to 5000cm^-1 and its spectral resolution could be chosen among six selected values 1, 2, 4, 8, 16, 32cm^-1. This paper will address the issues of image-guided spectroscopy and will show how an instrument designed for specifically imaging applications can dramatically improve the performance of the system and quality of the data acquired. The benefit of these technologies in spectroscopy can be demonstrated with a system optimally designed for detecting spectral characteristics of moving targets.