This PDF file contains the front matter associated with SPIE Proceedings Volume 10768, including the Title Page, Table of Contents, Author and Conference Committee lists.

We report an approximate simplified calculation of transmission spectra for a telecentric cone of light impinging on a Fabry-Perot interferometer. We model sulfur dioxide sensing and show that the F-number affects the optimum parameters. OCIS codes: (120.2230) Fabry-Perot; (280.1120) Air pollution monitoring

Detection of gases by a remote sensor has wide-ranging utility. The optimum sensor design depends on several key characteristics of the scene and imaging geometry. A generalized analysis of the signal-to-noise ratio (SNR) and spatial resolution shows that the optimum sensor F-number depends on the spatial distribution of the gas. Detection of a broad area emission results in a different optimal sensor configuration than when the emission is localized. The optimum F-number also depends on the exposure settings. A parametric analysis shows that detection performance varies between F_#^-2 and F_#⁴. The results of this paper provide guidance for sensor designers to optimize gas detection by imaging spectrometers.

Compressive sensing shows promise for sensors that collect fewer samples than required by traditional Shannon-Nyquist sampling theory. Recent sensor designs for hyperspectral imaging encode light using spectral modulators such as spatial light modulators, liquid crystal phase retarders, and Fabry-Perot resonators. The hyperspectral imager consists of a filter array followed by a detector array. It encodes spectra with less measurements than the number of bands in the signal, making reconstruction an underdetermined problem. We propose a reconstruction algorithm for hyperspectral images encoded through spectral modulators. Our approach constrains pixels to be similar to their neighbors in space and wavelength, as natural images tend to vary smoothly, and it increases robustness to noise. It combines L₁ minimization in the wavelet domain to enforce sparsity and total variation in the image domain for smoothness. The alternating direction method of multipliers (ADMM) simplifies the optimization procedure. Our algorithm constrains encoded, compressed hyperspectral images to be smooth in their reconstruction, and we present simulation results to illustrate our technique. This work improves the reconstruction of hyperspectral images from encoded, multiplexed, and sparse measurements.

Accurate retrieval of surface emissivity from long-wave infrared (LWIR) hyperspectral imaging data is necessary for many scientific and defense applications. Emissivity estimation consists of two interwoven steps: atmospheric compensation (AC) and temperature-emissivity separation (TES). AC uses an atmospheric estimate to convert the at-aperture radiance to ground radiance. Using the ground radiance, TES produces a temperature and emissivity estimate. TES algorithms require an accurate atmospheric model, and assumes that emissivity spectra for solids are smooth, compared to atmospheric features. A high-resolution atmospheric model is band-averaged to the sensor's spectral response function (SRF). Characterization and maintenance of the SRF is difficult, and errors cause rough emissivity estimates. We propose a method where spectra with smooth reflective emissivities are used to correct errors from the SRF. In-Scene AC (ISAC) methods can be used to find accurate estimates of the band-averaged atmospheric upwelling and transmission, but not the downwelling radiance which is needed for TES. Typical TES methods use a model for the downwelling radiance and an assumed SRF, which will differ from the true SRF causing unnaturally rough emissivity estimates. While ISAC estimates include the true SRF it is difficult to separate the SRF from these measurements. Instead of estimating the SRF directly, our method uses smooth low emissivity materials to produce a correction for the downwelling radiance that matches the true band-averaged values. We demonstrate this technique using simulated data.

The optical form of an imaging spectrometer for an airborne or space-borne EO/IR payload is often selected based on evaluation of the performance and size, weight, power, and cost (SWaP-C) of candidate spectrometer forms, the most commonly selected being Offner-Chrisp and Dyson spectrometers. Although these trades consider spectrometer metrics, the implications to the complexity and performance of the overall, higher-level system including other EO/IR sensors, are not always considered and may be far from equivalent. The Selectable Magnification Reflective Triplet (SMaRT) spectrometer is presented (shown in Figure 1) which offers greater system flexibility and potentially reduced complexity of the higher-level system. The SMaRT is a single-pass, unfolded form of the reflective triplet (RT) spectrometer. It has an RT collimator with real exit pupil at a dispersive element followed by a separate RT imager and focal plane –the focal lengths for the collimator and imager portions being independently selected. This allows an EO/IR payload with a SMaRT spectrometer and other EO/IR sensor(s) to potentially share imaging optics at an f-number selected for optimal performance of one or more of the other EO/IR sensors (e.g., optimized to provide high resolution context imagery), while separately selecting the f-number of the imager portion of the SMaRT to achieve the desired irradiance and resolution at the spectrometer focal plane. A system level trade is presented evaluating system complexity for comparably performing systems with a SMaRT or Offner-Chrisp spectrometer and high resolution context camera.

We describe an in-scene technique for accurate wavelength calibration of our airborne MAHI (Mid-infrared Airborne Hyperspectral Imager) sensor using modelled atmospheric spectral features. MAHI operates in the 3.3 to 5.4 μm region with a spectral sampling of 3.3 nm. The new technique significantly improves the accuracy over a lab technique using plastic film spectral features. We demonstrate the technique’s performance against: poorly known spectral response function; error in the initial wavelength grid guess; and deviation of the measured pixel spectra from the reference atmospheric spectrum. Performance of the new calibration technique for a recent airborne campaign in the Los Angeles area is demonstrated.

The hyperspectral thermal emission spectrometer (HyTES) participated in NASA’s HyspIRI Volcano and Coral Reef Airborne Campaign. The goal was to generate important science and application research results that uniquely enabled HyspIRI-like data while taking advantage of the contiguous spectroscopic measurements. HyTES was one of 4 sensors onboard a high altitude ER-2 aircraft. Other sensors included the Portable Remote Imaging SpectroMeter (PRISM), AVIRIS-classic, and MODIS/ASTER (MASTER). The flight lines gave us numerous opportunities to image the active Kilauea volcano (including Kona wind conditions) as well as urban coastlines. In preparation for the campaign, a calibration was performed to confirm the saturation temperature of HyTES as well as its sensitivity to applicable gas detection: sulfur dioxide (SO2), Hydrogen sulfide (H2S), and the Hydrogen halides. The HyTES sensor resides in a super pod on the wing and experiences near zero atmosphere Engineering challenges were overcome to keep the sensor calibrated during flight.

This paper focuses on the calibration and verification of the DESIS (DLR Earth Sensing Imaging Spectrometer) detector for the VIS/NIR (VNIR) spectral range. DESIS is a hyperspectral Instrument for the international space station, developed from the German Aerospace Center (DLR) and operate by Teledyne Brown Engineering (TBE). TBE provides the MUSES platform, on which the DESIS instrument will be mounted. The primary goal of DESIS is to measure and analyse quantitative diagnostic parameters describing key processes on the Earth surface. The main components of the sensor, the detector and the focal plane, were examined and verified. This allows predictions about the future data quality. The verification and validation of components and the entire system is an important and challenging task. The verification of the detectors is necessary to describe the characteristics of the detector according to predetermined specifications. The quantities to be examined are e.g. the quantum efficiency, the linearity of the detector, the pixel response non-uniformity (PRNU) and the dark current noise. For this purpose, specially calibrated integrated spheres are used that allow traceability of the measured data. With these information, the future performance of the sensor can be estimated using simulations.

We introduce a portable hand-held hyperspectral imaging system for the functional diagnostics of skin and vascular system. Hyperspectral image analysis aided by artificial neural networks (ANN) allows to reconstruct major physiological parameters of human skin nearly in real-time. The developed device provides spatial distribution of blood volume fraction, oxygenation and melanin content within skin. Special attention has been paid on the system validation and calibration using specially developed skin mimicking phantoms with confirmed optical properties. The device was built on the basis of unique hyperspectral snapshot camera utilizing a micro Fabry-Perot filter providing real spectral response in each pixel (no interpolation is used in image formation). A broadband illumination unit combined with the camera is based on the fiber-optic illuminator providing uniform distribution of light intensity and utilizes halogen lamp. The specially developed ANN algorithm was used to perform the inverse problem solution for quantitative assessment of major parameters of skin based on the measured hyperspectral images. A set of diffuse reflectance spectra of human skin imitated by the Monte Carlo method developed in-house has been used extensively for the training of ANN. The volume fraction of blood, oxygen saturation, melanin content and thickness of the epidermal layer were used variable parameters in the utilized seven-layer Monte Carlo-based skin model. The total training set contained 45,198 spectra in the range of 505–800 nm simulated with a step of 5 nm. The developed imaging system has been successfully used to perform the occlusion test measurements with healthy volunteers.

Previous work has shown Raman spectroscopy, together with statistical modeling, is effective for real-time data acquisition of consumable sugar (glucose) and accumulating products (butyric acid, acetic acid, and butanol) in Clostridium acetobutylicum cultures. Developed partial-least squares (PLS) models were applied to both agitated and static cultures with the former showing preferred modeling parameter values (R²Y = 0.99 and Q²Y = 0.98). Model outputs were comparable to off-line analyzed data from traditional HPLC for new clostridial experimental data through cross-validation. In this study, a cell-free system is explored in which experimental data from HPLC analyzed data for reaction components is used to simulate an 'artificial' fermentation culture devoid of cell activity or enzymes. Immersion probe data is assumed to not account for cell presence or associated activity in the cultures. Raman spectra of specific reaction components: (i) glucose, (ii) butyric acid, (iii) acetic acid, and (iv) butanol, in specified proportions were acquired for corresponding time points. The acquired spectra, together with known concentrations of reaction components, were used to build new sets of PLS models. Original cell-containing models and new cell-free models were run concurrently on new C. acetobutylicum fermentations. Comparison of model output results, suggest better predictability (e.g. Q²Y of 0.98 > Q²Y of 0.79) and less error (RMSECV of 0.98 < RMSECV of 2.76) in butyric acid concentrations for cell-containing models.

There is a high demand to apply photonics-based techniques to day-to-day routine use by targeting food quality control, food safeties, biosecurity and other related applications. We present the results of study of freshness and spoilage of meat samples by using visible and near-infrared spectroscopy. Two different configurations of experiment capable of sensing diverse range of probing depths have been tested. A table-top spectrophotometer equipped with an integrating sphere was used in one setup providing a large light spot and two detectors to measure two different spectral ranges (400-1100 nm and 1100-1700 nm). Another experimental setup comprised a fiber-optic linear array coupled to a portable spectrophotometer (measurement range: 400-1100 nm) to increase the average probing depth. The freshness of meat samples was assessed as a decrease of oxymyoglobin content monitored by decreasing 580/560 nm absorbance ratio, while spoilage was revealed by changes of absorbance at 1200 nm (fat content), 1450 nm (water), 1525 nm and 1600 nm (proteins) over time. We found that the studied meat samples experienced a significant loss of freshness after 2.5 hours. This could be interpreted as a beginning of the spoilage process showing promise of the applied methodology for spoilage sensing and benefits of the portable approach. The current technique is capable to real-time non-destructive screening of meat samples in a wide range wavelengths. The results and data analysis are presented and discussed in frame of food safety application.

Fluorescence spectroscopy and absorption spectroscopy are common physical methods used for water quality monitoring and analysis. However, in terms of sensitivity and selectivity, the absorption spectroscopy is still inferior; limited categories of organic contaminants can emit fluorescence, which constrains the analytical range. Here in, a novel feature extraction method is proposed in conjoint analysis of fluorescence and absorption spectroscopy to predict the category of water contaminants. The three-dimensional fluorescence spectra and absorption spectra of eight typical substances were studied. We extracted the outline of every three-dimensional fluorescence spectrum along the emission wavelengths axis, and then transformed it into a wavenumber spectrum. The symmetry axis and Stokes shift between fluorescence emission peak and absorption peak in their wavenumber spectra were set as two features. Theoretically, they depend only on the molecular structures of different substances. Moreover, four integral parameters in different absorption spectral ranges corresponding to functional groups were introduced to expand the analytical coverage of diverse contaminants including some non-fluorescent substances. Furthermore, we conducted long-term monitoring of river water near a dyeing and printing plant to demonstrate the prediction potential of this method. As an early warning system, the rapid prediction results can provide guidance for more targeted and detailed analysis and treatment.

The goal of this research was to compare the noise between the current MSI (multispectral imaging) R-CHIVE (Rochester Cultural Heritage Imaging, Visualization, and Education) system with various HSI (hyperspectral imaging) systems. Because most archived images are visually analyzed by humanities scholars, human vision factors are considered as limiting factors in both spatial resolution and minimum detectable noise levels for possible use as system design requirements. A major part of the image modeling analysis focused on the spatial/spectral compromise inherent in HSI systems and the impact of typical broadband sources' inefficiency in the visible spectrum. The peak LED spectral irradiance was found to be 180× greater than the broadband peak irradiance, and more than 2000× that of the visible region. This resulted in an increased scan time from 8 minutes to 2 hours if spatial resolution is required to be maintained at current levels for similar bands. Instituting an acceptable loss ratio from the very high resolution MSI system resulted in peak SNR values being maintained in as short as 5 seconds disregarding readout time. Signal to noise ratio (SNR) decreased from 106 in the MSI system to a peak of 57 in the modeled HSI system, following the well depth, or analog dynamic range of the sensor, as expected in a photon-limited scene.

The hyperspectral imaging system (HIS) using a Fourier transform infrared (FTIR) spectrometer is one of the key technologies for detection and identification of chemical warfare agents (CWAs). Recently, various detection algorithms based on machine learning techniques have been studied. These algorithms are robust against performance degradation caused by noise signatures generated by FTIR instruments. However, interference signatures from background materials degrade detection performance. In this paper, we propose an efficient algorithm that uses a support vector machine (SVM) classifier to detect CWAs. In contrast to the conventional algorithms that use measured spectra to train the SVM classifier, the proposed algorithm trains the SVM classifier using CWA signatures obtained by removing background signatures from measured spectra. Therefore, the proposed algorithm is robust against the performance degradation induced by interference signatures from background materials. Experimental results verify that the algorithm can detect CWA clouds effectively.

MODTRAN models the molecular absorption for the entire 0 to 50,000 cm^-1 spectral range. Typically, radiative transfer models define distinct line-shape functions depending on the spectral region. This can produce spectral anomalies at the transitions. A 3-parameter GrossDoppler line-shape function is defined that provides a spectrally-universal model for computing molecular absorption.

We describe a new algorithm, QUAC-IR (QUick Atmospheric Correction in the InfraRed), for automated, fast, atmospheric correction of LWIR (Long Wavelength InfraRed) hyperspectral imagery (HSI) and multi-spectral imagery (MSI) in the ~7-14 mm spectral region. QUAC-IR is an in-scene based algorithm, similar to the widely used ISAC (In- Scene Atmospheric Correction) algorithm. It improves upon the ISAC approach in several key ways, including providing absolute, versus relative, sensor-to-ground transmittances and radiances, as well as an estimate of the atmospheric downwelling sky radiance. The latter is important for retrieving emissivity from a reflective (i.e., non-blackbody) pixel. The key aspect of QUAC-IR is that it explicitly searches for blackbody pixels using an efficient approach involving a small number of spectral channels in which the atmospheric radiative transfer is dominated by the water continuum. This allows for fast and simplified Beer's Law (i.e., exponential) scaling of the path transmittance and radiance based on a compact library of pre-computed reference values. We apply QUAC-IR to well-calibrated data from the SEABASS¹ and MAKO² HSI sensors. The results are compared to those from a first-principles physics-based atmospheric code, FLAASH-IR.

Optimal interpretation of remote sensing imagery requires characterizing the atmospheric composition between a sensor and the area it is observing. Timely estimates of atmospheric temperature, water vapor, and other constituents from the ground to the edge of the space environment are not always readily available. In those cases, we must supplement our knowledge of the atmosphere’s composition to fill in any gaps in knowledge and empirical models of the atmosphere are useful tools for this purpose. The Standardized Atmosphere Generator (SAG) was constructed is one such empirical. It has been designed to allow all the major known, systematic variability in the atmosphere and may be used to generate atmospheric profile from the ground to 300 km consistent with user-specified temporal, geophysical, and geographical information Output provides reasonable estimates for temperature, pressure, and densities of atmospheric constituents and can be directly incorporated into radiative transfer forward models or retrieval algorithms. SAG draws upon a number of existing empirical atmospheric models and ensures consistency of output between them. It can be used either as a stand-alone interactive program or scripted for batch execution and assist in determining atmospheric attenuation, refraction, scattering, chemical kinetic temperature profiles, and a host of other naturally occurring processes. Here, we will discuss the capabilities and performance of the SAG model for a variety of applications including its interactive and batch processing use. We will also demonstrate the physical realism of SAG through a small number of relevant use cases.

This paper presents an UltraViolet-Visible (UV-Vis) spectral radiance simulation capability for Non-Local Thermodynamic Equilibrium (non-LTE) conditions, consisting of a full line-by-line (LBL) radiative transfer (RT) algorithm and a UV-Vis signatures library. Results are presented for two example scenarios where strong UV-Vis emissions arise, an atmospheric high altitude auroral event and a High Explosive (HE) detonation.

Hyperspectral instruments are fundamental tools in remote sensing for environmental control and precision farming. For hyperspectral sensors often conventional optical designs based on grating or prism spectrometers are preferred. These instruments meet the system and mission requirements. From the provided data high quality information can be derived. However, more and more data is being offered by low-cost missions. This will establish new business models and data providers. This article is intended to provide an overview of current low cost sensors for hyperspectral applications.

We propose to utilize spectral information of photoluminescence (PL) photons for position sensing. Suppose that a beam of radiation is incident at a certain point on a rectangular plate in which luminescent materials are uniformly dispersed. With an optical fiber attached to its edge surface, the PL photons are guided to a spectrometer. The spectrum of the PL photons is red-shifted due to self-absorption in the plate. The magnitude of the red-shift is enhanced as the PL photons propagate longer distance inside the plate. Hence, we can determine the distance by quantifying this spectral change. In experiment, we let a laser beam (wavelength 450nm) normally incident on an acrylic plate containing luminescent materials (40 mm × 40 mm × 2.9 mm). The incident position on the plate was varied and from each spectrum recorded we calculated chromaticity coordinates in the CIE1931-XYZ color space. With one sample plate, the coordinate x increased from 0.23 to 0.29 monotonically when we increased the horizontal distance on the plate from 2mm to 20mm. In another sample, the chromaticity coordinates behaved differently but the monotonic relation remained valid. We now have calibration curves for the position. This sensing technique might be suited for long-range position detection, usage in harsh environments and for insertion to narrow places.

We have demonstrated one-dimensional position sensing based on red-shift of photoluminescence (PL) spectra. This technology can be extended to two-dimension by the two-stage PL conversion technique described here. A planar waveguide contains a first luminophore to convert an incoming radiation to PL photons. A linear waveguide contains a second luminophore and two of them sandwich the planar waveguide such that the second luminophore absorbs the PL photons emitted by the first luminophore. Spectral analysis of the PL photons exiting the two linear waveguides gives the coordinate of the incident position along the direction of the linear waveguide. The coordinate perpendicular to this direction is determined by comparing the PL intensities propagating in the two linear waveguides. This is analogous to the charge division principle utilized in a position-sensitive proportional counter as well as a tetra-lateral semiconductor detector. In the current case, we are dividing the PL photons emitted by the first luminophore to the two light-sensitive regions facing to each other. The use of optical fibers allows one to build optical sensors without electric components at the sensing sites. Based on this technology, a robust large-scale radiation monitoring system might be constructed. In a proof-of-concept experiment, we fabricated a 50 × 50 × 8 mm sensor head using coumarin6 as a green emitter and Lumogen F Red 305 as a red emitter. The maximum error for estimating the incident spots in the upper-left 20 × 20 mm region of the sensor head was 1.5mm.