Thermal Hyperspectral Imaging Research Articles

Applications of quantitative thermal infrared hyperspectral imaging (8–14 μm): measuring volcanic SO2 mass flux and determining plume transport velocity using a single sensor

Remote measurements of sulfur dioxide (SO2) path-concentration and emission rate (mass flux), released by passive volcanic degassing, are a key diagnostic of volcanic behavior. SO2 gas concentrations are also important for public health. Despite the significance of measuring SO2 mass flux at active volcanoes, an accurate method is still very difficult to obtain due largely to uncertainties of plume transport velocities and radiative transfer calculations. Here, we present a new infrared imaging method for deriving SO2 flux at active volcanoes with less than 20% plume transport velocity and radiative transfer uncertainties. The thermal hyperspectral imager (THI) was used for this study. THI is an uncooled remote sensing long-wave thermal infrared (TIR) imaging hyperspectral sensor that can obtain ~ 50 wavelength samples between 8 and 14 μm. Measurements of SO2 at the summit of Kīlauea volcano, collected using THI, are presented, where we obtained (for the periods of July 24–26, 2017, and February 5–6, 2018) a SO2 flux raging between 0 and 20 kg/s with occasional peaks higher than 40 kg/s. The spatial distribution of SO2 path-concentrations was obtained from the THI images by processing them using a newly developed SO2 retrieval algorithm, the SO2 amenable lookup table algorithm (SO2-ALTA). The sampling rate (30 Hz) of the microbolometer camera within THI means raw frames can be used for wind velocity derivation. Volcanic plume motion can be thus inferred by tracking plume features in sequential frames using spatial correlation techniques and knowing the camera angular orientation. Volcanic SO2 flux was estimated from the spatial distribution of SO2 and wind velocity. We evaluated temporal trends of the SO2 flux during the day and the night at the summit of Kīlauea volcano under both clear sky weather with northeasterly trade winds and cloudy sky conditions with low variable wind. These tests showed that the camera and techniques described in this paper provide an effective tool for monitoring SO2 fluxes remotely with less than 20% plume transport velocity and radiative transfer uncertainties and under a variety of background conditions (clear and cloudy sky, cold and hot backgrounds).

Developed split window method to retrieve land surface temperature from HYTES thermal hyperspectral images using genetic algorithm

ABSTRACTThis article aims to establish a new method to retrieve land surface temperature (LST) from hyperspectral thermal emission spectrometer (HYTES) data with split window (SW) algorithm. First, the optimal bands of HYTES sensor were selected with the genetic algorithm and then were used in the SW algorithm. In the SW algorithm, its coefficients were obtained based on several subranges of atmospheric column water vapours (CWVs) and view zenith angle (VZA) under various land surface conditions, in order to remove the atmospheric effect and improve the retrieval accuracy. Results showed that the root-mean-square error (RMSE) varies for different CWV and VZA, and with the increasing CWV and VZA, the RMSE value also increases. The emissivity, CWV, and VZA were also obtained for pixels. The sensitive analysis of LST retrieval to instrument noise and uncertainty of pixel emissivity and water vapour demonstrated the good performance of the proposed algorithm. Finally, the new algorithm was applied to HYTES sensor data, and the LST was validated using LST product of HYTES sensor obtained by NASA. The results showed that the RMSE of the LST retrieval with the proposed algorithm and the LST product of sensor for data 1 and data 2 is 1.3 K and 1.6 K, respectively.

Improved Alpha Residuals for Target Detection in Thermal Hyperspectral Imaging

This letter presents a new method to compute alpha residuals (AR) in longwave infrared hyperspectral images. AR allow emissivity-based target detection being them a quantity, related to the target emissivity, which is less prone to errors in the estimated temperature. The proposed method improves existing approaches taking into account also the contribution of the atmospheric emitted radiance reflected by the object. Experimental results over synthetic data show the better performance of the new approach.

Toward Optimum Fusion of Thermal Hyperspectral and Visible Images in Classification of Urban Area

Recently, classification of urban area based on multi-sensor fusion has been widely investigated. In this paper, the potential of using visible (VIS) and thermal infrared (TIR) hyperspectral images fusion for classification of urban area is evaluated. For this purpose, comprehensive spatial-spectral feature space is generated which includes vegetation index, differential morphological profile (DMP), attribute profile (AP), texture, geostatistical features, structural feature set (SFS) and local statistical descriptors from both datasets in addition to original datasets. Although Support Vector Machine (SVM) is an appropriate tool in the classification of high dimensional feature space, its performance is significantly affected by its parameters and feature space. Cuckoo search (CS) optimization algorithm with mixed binary-continuous coding is proposed for feature selection and SVM parameter determination simultaneously. Moreover, the significance of each selected feature category in the classification of a specific object is verified. Accuracy assessment on two subsets shows that stacking of VIS and TIR bands can improve the classification performance to 87 percent and 82 percent for two subsets, compare to VIS image (72 percent and 80 percent) and TIR image (50 percent and 56 percent). However, the optimum results obtained based on the proposed method which gains 94 percent and 92 percent. Furthermore, results show that using TIR beside VIS image improves classification accuracy of roads and buildings in urban area.

Toward Optimum Fusion of Thermal Hyperspectral and Visible Images in Classification of Urban Area

Recently, classification of urban area based on multi-sensor fusion has been widely investigated. In this paper, the potential of using visible (<small>VIS</small>) and thermal infrared (<small>TIR</small>) hyperspectral images fusion for classification of urban area is evaluated. For this purpose, comprehensive spatial-spectral feature space is generated which includes vegetation index, differential morphological profile (<small>DMP</small>), attribute profile (<small>AP</small>), texture, geostatistical features, structural feature set (<small>SFS</small>) and local statistical descriptors from both datasets in addition to original datasets. Although Support Vector Machine (<small>SVM</small>) is an appropriate tool in the classification of high dimensional feature space, its performance is significantly affected by its parameters and feature space. Cuckoo search (<small>CS</small>) optimization algorithm with mixed binary-continuous coding is proposed for feature selection and <small>SVM</small> parameter determination simultaneously. Moreover, the significance of each selected feature category in the classification of a specific object is verified. Accuracy assessment on two subsets shows that stacking of <small>VIS</small> and <small>TIR</small> bands can improve the classification performance to 87 percent and 82 percent for two subsets, compare to <small>VIS</small> image (72 percent and 80 percent) and <small>TIR</small> image (50 percent and 56 percent). However, the optimum results obtained based on the proposed method which gains 94 percent and 92 percent. Furthermore, results show that using <small>TIR</small> beside <small>VIS</small> image improves classification accuracy of roads and buildings in urban area.

Texture-Guided Multisensor Superresolution for Remotely Sensed Images

This paper presents a novel technique, namely texture-guided multisensor superresolution (TGMS), for fusing a pair of multisensor multiresolution images to enhance the spatial resolution of a lower-resolution data source. TGMS is based on multiresolution analysis, taking object structures and image textures in the higher-resolution image into consideration. TGMS is designed to be robust against misregistration and the resolution ratio and applicable to a wide variety of multisensor superresolution problems in remote sensing. The proposed methodology is applied to six different types of multisensor superresolution, which fuse the following image pairs: multispectral and panchromatic images, hyperspectral and panchromatic images, hyperspectral and multispectral images, optical and synthetic aperture radar images, thermal-hyperspectral and RGB images, and digital elevation model and multispectral images. The experimental results demonstrate the effectiveness and high general versatility of TGMS.

Optimizing Object, Atmosphere, and Sensor Parameters in Thermal Hyperspectral Imagery

We address the problem of estimating atmosphere parameters (temperature and water vapor content) from data captured by an airborne thermal hyperspectral imager and propose a method based on linear and nonlinear optimization. The method is used for the estimation of the parameters (temperature and emissivity) of the observed object as well as sensor gain under certain restrictions. The method is analyzed with respect to sensitivity to noise and the number of spectral bands. Simulations with synthetic signatures are performed to validate the analysis, showing that the estimation can be performed with as few as 10–20 spectral bands at moderate noise levels. The proposed method is also extended to exploit additional knowledge, for example, measurements of atmospheric parameters and sensor noise. Additionally, we show how to extend the method in order to improve spectral calibration.

Characterization and initial field test of an 8–14 μm thermal infrared hyperspectral imager for measuring SO2 in volcanic plumes

The ability to image and quantify SO2 path-concentrations in volcanic plumes, either by day or by night, is beneficial to volcanologists. Gas emission rates are affected by the chemical equilibria in rising magmas and a better understanding of this relationship would be useful for short-term eruption prediction. A newly developed remote sensing long-wave thermal InfraRed (IR) imaging hyperspectral sensor – the Thermal Hyperspectral Imager (THI) – was built and tested. The system employs a Sagnac interferometer and an uncooled microbolometer in rapid scanning configuration to collect hyperspectral images of volcanic plumes. Each pixel in the resulting image yields a spectrum with 50 samples between 8 and 14 μm. Images are spectrally and radiometrically calibrated using an IR source with a narrow band filter and two blackbodies. In this paper, the sensitivity of the instrument for the purpose of quantifying SO2 using well constrained laboratory experiments is evaluated, and initial field results from Kīlauea volcano, Hawai’i, are presented. The sensitivity of THI was determined using gas cells filled with known concentrations of SO2 and using NIST-traceable blackbodies to simulate a range of realistic background conditions. Measurements made by THI were then benchmarked against a high spectral resolution off-the-shelf Michelson FTIR instrument. Theoretical thermal IR spectral radiances were computed with MODTRAN5 for the same optical conditions, to evaluate how well the (known) concentration of SO2 in the gas cells could be retrieved from the resulting THI spectra. Finally, THI was recently field-tested at Kīlauea to evaluate its ability to image the concentration of SO2 in a real volcanic plume. A path-concentration of 7150 ppm m was retrieved from measurements made near the Halema’uma’u vent.

Space-based hyperspectral technologies for the thermal infrared

Various approaches now exist for obtaining spectral imagery over a broad range of infrared wavelengths. One approach involves the use of a single grating element in two grating orders with dualband focal plane array (FPA) technology. This approach offers greater efficiency over the mid-wave infrared and long-wave infrared and eliminates the need for separate FPAs, dispersing elements, and optical beamsplitters. Another approach achieves similar results by exploiting an FPA which has a broad wavelength response with an innovative grating which has useable efficiency that extends beyond the single octave limits of traditional gratings. Significant advantages result, in either case, for space-based hyperspectral imagers, for which a reduction in cryo-cooled mass translates into prodigious savings in overall payload mass, cryo-cooling requirements, and waste heat removal. By contrast, longer term approaches might realize infrared “hyperspectral pixels” in two-dimensional imaging FPAs. In this case, each pixel would detect different wavelengths of radiation, at different depths, and the resulting “spectral photocurrents” would be transported to read-out circuitry through a vertical grid of electrical contacts. Although not yet realized in practice, the conceptual basis for accomplishing this, with the widely available HgCdTe detector material, has been described. With regard to employment, space-based thermal hyperspectral imaging is characterized by coarser ground resolution as a result of aperture diameter limitations and diffraction considerations at the longer infrared wavelengths. The resulting subpixel detections, based on spectral signature, are often complementary with higher resolution, shorter wavelength, panchromatic imagery.

BBM/EM design of the thermal hyperspectral imager: An instrument for remote sensing of earth's surface, atmosphere and ocean, from a microsatellite platform

The Thermal Hyperspectral Imager (THI) is a low cost, low mass, power efficient instrument designed to acquire hyperspectral remote sensing data in the long-wave infrared. The instrument has been designed to satisfy mass, volume, and power constraints necessary to allow for its accommodation in a 95kg micro-satellite bus, designed by staff and students at the University of Hawai'i. THI acquires approximately 30 separate spectral bands in the 8–14μm wavelength region, at 16 wavenumber resolution. Rather than using filtering or dispersion to generate the spectral information, THI uses an interferometric technique. Light from the scene is focused onto an uncooled microbolometer detector array through a stationary interferometer, causing the light incident at each detector at any instant in time to be phase shifted by an optical path difference which varies linearly across the array in the along-track dimension. As platform motion translates the detector array in the along-track direction at a rate of approximately one pixel per frame (the camera acquires data at 30Hz) the radiance from each scene element can be sampled at each OPD, thus generating an interferogram. Spectral radiance as a function of wavelength is subsequently obtained for each scene element using standard Fourier transform techniques. Housed in a pressure vessel to shield COTS parts from the space environment, the total instrument has a mass of 15kg. Peak power consumption, largely associated with the calibration procedure, is <90W. From a nominal altitude of 550km the resulting data would have a spatial resolution of approximately 300m. Although an individual imaging event yields approximately 1Gbit of raw uncompressed data, onboard processing (to convert the interferograms into a conventional spectral hypercube) can reduce this to tens of Mega bits per scene. In this presentation we will describe (a) the rationale for the project, (b) the instrument design, and (c) how the data are processed. Finally we will present data acquired by THI on a laboratory microscope stage to demonstrate the spectro-radiometric quality of the data that the instrument can provide.

Thermal hyperspectral imagers and their applications

Thermal hyperspectral imagers and their applications

Quantum well earth science testbed

A thermal hyperspectral imager is underdevelopment which utilizes the compact Dyson optical configuration and the broadband (8–12 μm) quantum well infrared photodetector (QWIP) focal plane array technology. The Dyson configuration uses a single monolithic prism-like grating design which allows for a high throughput instrument (F/1.6) with minimal ghosting, stray light and large swath width. The configuration has the potential to be the optimal high resolution imaging spectroscopy solution for aerial and space remote sensing applications due to its small form factor and relatively low power requirements. The planned instrument specifications are discussed as well as thermal design trade-offs. The current design uses a single high power cryocooler which allows operation of the QWIP at 40 K with adequate temperature stability. Calibration testing results (noise equivalent temperature difference, spectral linearity and spectral bandwidth) and laboratory emissivity plots from samples are shown using an operational testbed unit which has similar specifications as the final airborne system. Field testing of the testbed unit was performed to acquire plots of emissivity for various known standard minerals (quartz, opal, alunite). A comparison is made using data from the ASTER spectral library. The current single band (8–9 μm) testbed utilizes the high uniformity and operability of the QWIP array and shows excellent laboratory and field spectroscopic results.