SEARCH · Engineering Papers
Search indexed NASA NTRS and DOE OSTI research on propulsion, heat transfer, battery materials and energy systems. Follow report and document links to the original sources.
Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.
HICO and RAIDS Experiment Payload - Hyperspectral Imager For The Coastal Ocean (HREP-HICO) will operate a visible and near-infrared (VNIR) Maritime Hyperspectral Imaging (MHSI) system, to detect, identify and quantify coastal geophysical features from the International Space Station.
Recent development of a polarimetric hyperspectral imaging (PHSI) prototype sensor system that uses an acousto-optic tunable filter (AOTF) as a wavelength sorter and a polarizing beam splitter.
High-speed, low-power, reconfigurable electronic hardware has been developed to implement ICER-3D, an algorithm for compressing hyperspectral-image data. The algorithm and parts thereof have been the topics of several NASA Tech Briefs articles, including Context Modeler for Wavelet Compression of Hyperspectral Images (NPO-43239) and ICER-3D Hyperspectral Image Compression Software (NPO-43238), which appear elsewhere in this issue of NASA Tech Briefs. As described in more detail in those articles, the algorithm includes three main subalgorithms: one for computing wavelet transforms, one for context modeling, and one for entropy encoding. For the purpose of designing the hardware, these subalgorithms are treated as modules to be implemented efficiently in field-programmable gate arrays (FPGAs). The design takes advantage of industry- standard, commercially available FPGAs. The implementation targets the Xilinx Virtex II pro architecture, which has embedded PowerPC processor cores with flexible on-chip bus architecture. It incorporates an efficient parallel and pipelined architecture to compress the three-dimensional image data. The design provides for internal buffering to minimize intensive input/output operations while making efficient use of offchip memory. The design is scalable in that the subalgorithms are implemented as independent hardware modules that can be combined in parallel to increase throughput. The on-chip processor manages the overall operation of the compression system, including execution of the top-level control functions as well as scheduling, initiating, and monitoring processes. The design prototype has been demonstrated to be capable of compressing hyperspectral data at a rate of 4.5 megasamples per second at a conservative clock frequency of 50 MHz, with a potential for substantially greater throughput at a higher clock frequency. The power consumption of the prototype is less than 6.5 W. The reconfigurability (by means of reprogramming) of the FPGAs makes it possible to effectively alter the design to some extent to satisfy different requirements without adding hardware. The implementation could be easily propagated to future FPGA generations and/or to custom application-specific integrated circuits.
A new spectral-spatial method for hyperspectral data classification is proposed. For a given hyperspectral image, probabilistic pixelwise classification is first applied. Then, hierarchical step-wise optimization algorithm is performed, by iteratively merging neighboring regions with the smallest Dissimilarity Criterion (DC) and recomputing class labels for new regions. The DC is computed by comparing region mean vectors, class labels and a number of pixels in the two regions under consideration. The algorithm is converged when all the pixels get involved in the region merging procedure. Experimental results are presented on two remote sensing hyperspectral images acquired by the AVIRIS and ROSIS sensors. The proposed approach improves classification accuracies and provides maps with more homogeneous regions, when compared to previously proposed classification techniques.
A long-wave infrared hyperspectral sensor device employs a combination of an interferometer with an uncooled microbolometer array camera to produce hyperspectral images without the use of bulky, power-hungry motorized components, making it suitable for UAV vehicles, small mobile platforms, or in extraterrestrial environments. The sensor device can provide signal-to-noise ratios near 200 for ambient temperature scenes with 33 wavenumber resolution at a frame rate of 50 Hz, with higher results indicated by ongoing component improvements.
A software package generates simulated hyperspectral imagery for use in validating algorithms that generate estimates of Earth-surface spectral reflectance from hyperspectral images acquired by airborne and spaceborne instruments. This software is based on a direct simulation Monte Carlo approach for modeling three-dimensional atmospheric radiative transport, as well as reflections from surfaces characterized by spatially inhomogeneous bidirectional reflectance distribution functions. In this approach, "ground truth" is accurately known through input specification of surface and atmospheric properties, and it is practical to consider wide variations of these properties. The software can treat both land and ocean surfaces, as well as the effects of finite clouds with surface shadowing. The spectral/spatial data cubes computed by use of this software can serve both as a substitute for, and a supplement to, field validation data.
Software was developed that automatically detects minerals that are present in each pixel of a hyperspectral image. An algorithm based on sparse spectral unmixing with Bayesian Positive Source Separation is used to produce mineral abundance maps from hyperspectral images. A superpixel segmentation strategy enables efficient unmixing in an interactive session. The algorithm computes statistically likely combinations of constituents based on a set of possible constituent minerals whose abundances are uncertain. A library of source spectra from laboratory experiments or previous remote observations is used. A superpixel segmentation strategy improves analysis time by orders of magnitude, permitting incorporation into an interactive user session (see figure). Mineralogical search strategies can be categorized as supervised or unsupervised. Supervised methods use a detection function, developed on previous data by hand or statistical techniques, to identify one or more specific target signals. Purely unsupervised results are not always physically meaningful, and may ignore subtle or localized mineralogy since they aim to minimize reconstruction error over the entire image. This algorithm offers advantages of both methods, providing meaningful physical interpretations and sensitivity to subtle or unexpected minerals.
Explore the source record for details and available documents.
Software has been developed to implement the ICER-3D algorithm. ICER-3D effects progressive, three-dimensional (3D), wavelet-based compression of hyperspectral images. If a compressed data stream is truncated, the progressive nature of the algorithm enables reconstruction of hyperspectral data at fidelity commensurate with the given data volume. The ICER-3D software is capable of providing either lossless or lossy compression, and incorporates an error-containment scheme to limit the effects of data loss during transmission. The compression algorithm, which was derived from the ICER image compression algorithm, includes wavelet-transform, context-modeling, and entropy coding subalgorithms. The 3D wavelet decomposition structure used by ICER-3D exploits correlations in all three dimensions of sets of hyperspectral image data, while facilitating elimination of spectral ringing artifacts, using a technique summarized in "Improving 3D Wavelet-Based Compression of Spectral Images" (NPO-41381), NASA Tech Briefs, Vol. 33, No. 3 (March 2009), page 7a. Correlation is further exploited by a context-modeling subalgorithm, which exploits spectral dependencies in the wavelet-transformed hyperspectral data, using an algorithm that is summarized in "Context Modeler for Wavelet Compression of Hyperspectral Images" (NPO-43239), which follows this article. An important feature of ICER-3D is a scheme for limiting the adverse effects of loss of data during transmission. In this scheme, as in the similar scheme used by ICER, the spatial-frequency domain is partitioned into rectangular error-containment regions. In ICER-3D, the partitions extend through all the wavelength bands. The data in each partition are compressed independently of those in the other partitions, so that loss or corruption of data from any partition does not affect the other partitions. Furthermore, because compression is progressive within each partition, when data are lost, any data from that partition received prior to the loss can be used to reconstruct that partition at lower fidelity. By virtue of the compression improvement it achieves relative to previous means of onboard data compression, this software enables (1) increased return of hyperspectral scientific data in the presence of limits on the rates of transmission of data from spacecraft to Earth via radio communication links and/or (2) reduction in spacecraft radio-communication power and/or cost through reduction in the amounts of data required to be downlinked and stored onboard prior to downlink. The software is also suitable for compressing hyperspectral images for ground storage or archival purposes.
The proposal requested funds for the computing power to bring hyperspectral image processing into undergraduate and graduate remote sensing courses. This upgrade made it possible to handle more students in these oversubscribed courses and to enhance CSES' summer short course entitled "Hyperspectral Imaging and Data Analysis" provided for government, industry, university and military. Funds were also requested to build field measurement capabilities through the purchase of spectroradiometers, canopy radiation sensors and a differential GPS system. These instruments provided systematic and complete sets of field data for the analysis of hyperspectral data with the appropriate radiometric and wavelength calibration as well as atmospheric data needed for application of radiative transfer models. The proposed field equipment made it possible to team-teach a new field methods course, unique in the country, that took advantage of the expertise of the investigators rostered in three different departments, Geology, Geography and Biology.
A hyperspectral imager includes a focal plane having an array of spaced image recording pixels receiving light from a scene moving relative to the focal plane in a longitudinal direction, the recording pixels being transportable at a controllable rate in the focal plane in the longitudinal direction, an electronic shutter for adjusting an exposure time of the focal plane, whereby recording pixels in an active area of the focal plane are removed therefrom and stored upon expiration of the exposure time, an electronic spectral filter for selecting a spectral band of light received by the focal plane from the scene during each exposure time and an electronic controller connected to the focal plane, to the electronic shutter and to the electronic spectral filter for controlling (1) the controllable rate at which the recording is transported in the longitudinal direction, (2) the exposure time, and (3) the spectral band so as to record a selected portion of the scene through M spectral bands with a respective exposure time t(sub q) for each respective spectral band q.
JPL developed a prototype system for evaluating potential applications of a polarimetric hyperspectral imaging (PHI) technology using the noncollinear acousto-optic tunable filter (AOTF).
Hyperspectral data are becoming more widely available via sensors on airborne and unmanned aerial vehicle (UAV) platforms, as well as proximal platforms. While space-based hyperspectral data continue to be limited in availability, multiple spaceborne Earth-observing missions on traditional platforms are scheduled for launch, and companies are experimenting with small satellites for constellations to observe the Earth, as well as for planetary missions. Land cover mapping via classification is one of the most important applications of hyperspectral remote sensing and will increase in significance as time series of imagery are more readily available. However, while the narrow bands of hyperspectral data provide new opportunities for chemistry-based modeling and mapping, challenges remain. Hyperspectral data are high dimensional, and many bands are highly correlated or irrelevant for a given classification problem. For supervised classification methods, the quantity of training data is typically limited relative to the dimension of the input space. The resulting Hughes phenomenon, often referred to as the curse of dimensionality, increases potential for unstable parameter estimates, overfitting, and poor generalization of classifiers. This is particularly problematic for parametric approaches such as Gaussian maximum likelihood–based classifiers that have been the backbone of pixel-based multispectral classification methods. This issue has motivated investigation of alternatives, including regularization of the class covariance matrices, ensembles of weak classifiers, development of feature selection and extraction methods, adoption of nonparametric classifiers, and exploration of methods to exploit unlabeled samples via semi-supervised and active learning. Data sets are also quite large, motivating computationally efficient algorithms and implementations. This chapter provides an overview of the recent advances in classification methods for mapping vegetation using hyperspectral data. Three data sets that are used in the hyperspectral classification literature (e.g., Botswana Hyperion satellite data and AVIRIS airborne data over both Kennedy Space Center and Indian Pines) are described in Section 3.2 and used to illustrate methods described in the chapter. An additional high-resolution hyperspectral data set acquired by a SpecTIR sensor on an airborne platform over the Indian Pines area is included to exemplify the use of new deep learning approaches, and a multiplatform example of airborne hyperspectral data is provided to demonstrate transfer learning in hyperspectral image classification. Classical approaches for supervised and unsupervised feature selection and extraction are reviewed in Section 3.3. In particular, nonlinearities exhibited in hyperspectral imagery have motivated development of nonlinear feature extraction methods in manifold learning, which are outlined in Section 3.3.1.4. Spatial context is also important in classification of both natural vegetation with complex textural patterns and large agricultural fields with significant local variability within fields. Approaches to exploit spatial features at both the pixel level (e.g., co-occurrence–based texture and extended morphological attribute profiles [EMAPs]) and integration of segmentation approaches (e.g., HSeg) are discussed in this context in Section 3.3.2. Recently, classification methods that leverage nonparametric methods originating in the machine learning community have grown in popularity. An overview of both widely used and newly emerging approaches, including support vector machines (SVMs), Gaussian mixture models, and deep learning based on convolutional neural networks is provided in Section 3.4. Strategies to exploit unlabeled samples, including active learning and metric learning, which combine feature extraction and augmentation of the pool of training samples in an active learning framework, are outlined in Section 3.5. Integration of image segmentation with classification to accommodate spatial coherence typically observed in vegetation is also explored, including as an integrated active learning system. Exploitation of multisensor strategies for augmenting the pool of training samples is investigated via a transfer learning framework in Section 3.5.1.2. Finally, we look to the future, considering opportunities soon to be provided by new paradigms, as hyperspectral sensing is becoming common at multiple scales from ground-based and airborne autonomous vehicles to manned aircraft and space-based platforms.
A new spectral-spatial method for classification of hyperspectral images is proposed. The HSegClas method is based on the integration of probabilistic classification and shape analysis within the hierarchical step-wise optimization algorithm. First, probabilistic support vector machines classification is applied. Then, at each iteration two neighboring regions with the smallest Dissimilarity Criterion (DC) are merged, and classification probabilities are recomputed. The important contribution of this work consists in estimating a DC between regions as a function of statistical, classification and geometrical (area and rectangularity) features. Experimental results are presented on a 102-band ROSIS image of the Center of Pavia, Italy. The developed approach yields more accurate classification results when compared to previously proposed methods.
This paper reports activities in the development of AOTF Polarimetric Hyperspectral Imaging (PHI) Systems at JPL along with field observation results for illustrating the technology capabilities and advantages in remote sensing. In addition, the technology was also used to measure thickness distribution and structural imperfections of silicon-on-silicon wafers using white light interference phenomenon for demonstrating the potential in scientific and industrial applications.
The characterization and calibration of hyperspectral imagers is a challenging one that is expected to become even more challenging as needs increase for highly-accurate radiometric data from such systems. The preflight calibration of the Advanced Responsive Tactically Effective Military Imaging Spectrometer (ARTEMIS) is used as an example of the difficulties to calibrate hyperspectrally. Results from a preflight solar radiation-based calibration are presented with a discussion of the uncertainties in such a method including the NISI-traceable and SItraceable aspects. Expansion on the concept of solar-based calibration is given with descriptions of methods that view the solar disk directly, illuminate a solar diffuser that is part of the sensor's inflight calibration, and illuminate an external diffuser that is imaged by the sensor. The results of error analysis show that it is feasible to achieve preflight calibration using the sun as a source at the same level of uncertainty as those of lamp-based approaches. The error analysis is evaluated and verified through the solar-radiation-based calibration of several of laboratory grade radiometers. Application of these approaches to NASA's upcoming CLARREO mission are discussed including proposed methods for significantly reducing the uncertainties to allow CLARREO data to be used for climate data records.
Efficient onboard data compression can reduce the data volume from hyperspectral imagers on NASA and DoD spacecraft in order to return as much imagery as possible through constrained downlink channels. Lossless compression is important for signature extraction, object recognition, and feature classification capabilities. To provide onboard data compression, a hardware implementation of a lossless hyperspectral compression algorithm was developed using a field programmable gate array (FPGA). The underlying algorithm is the Fast Lossless (FL) compression algorithm reported in Fast Lossless Compression of Multispectral- Image Data (NPO-42517), NASA Tech Briefs, Vol. 30, No. 8 (August 2006), p. 26 with the modification reported in Lossless, Multi-Spectral Data Comressor for Improved Compression for Pushbroom-Type Instruments (NPO-45473), NASA Tech Briefs, Vol. 32, No. 7 (July 2008) p. 63, which provides improved compression performance for data from pushbroom-type imagers. An FPGA implementation of the unmodified FL algorithm was previously developed and reported in Fast and Adaptive Lossless Onboard Hyperspectral Data Compression System (NPO-46867), NASA Tech Briefs, Vol. 36, No. 5 (May 2012) p. 42. The essence of the FL algorithm is adaptive linear predictive compression using the sign algorithm for filter adaption. The FL compressor achieves a combination of low complexity and compression effectiveness that exceeds that of stateof- the-art techniques currently in use. The modification changes the predictor structure to tolerate differences in sensitivity of different detector elements, as occurs in pushbroom-type imagers, which are suitable for spacecraft use. The FPGA implementation offers a low-cost, flexible solution compared to traditional ASIC (application specific integrated circuit) and can be integrated as an intellectual property (IP) for part of, e.g., a design that manages the instrument interface. The FPGA implementation was benchmarked on the Xilinx Virtex IV LX25 device, and ported to a Xilinx prototype board. The current implementation has a critical path of 29.5 ns, which dictated a clock speed of 33 MHz. The critical path delay is end-to-end measurement between the uncompressed input data and the output compression data stream. The implementation compresses one sample every clock cycle, which results in a speed of 33 Msample/s. The implementation has a rather low device use of the Xilinx Virtex IV LX25, making the total power consumption of the implementation about 1.27 W.
Two methods of increasing the effectiveness of three-dimensional (3D) wavelet-based compression of hyperspectral images have been developed. (As used here, images signifies both images and digital data representing images.) The methods are oriented toward reducing or eliminating detrimental effects of a phenomenon, referred to as spectral ringing, that is described below. In 3D wavelet-based compression, an image is represented by a multiresolution wavelet decomposition consisting of several subbands obtained by applying wavelet transforms in the two spatial dimensions corresponding to the two spatial coordinate axes of the image plane, and by applying wavelet transforms in the spectral dimension. Spectral ringing is named after the more familiar spatial ringing (spurious spatial oscillations) that can be seen parallel to and near edges in ordinary images reconstructed from compressed data. These ringing phenomena are attributable to effects of quantization. In hyperspectral data, the individual spectral bands play the role of edges, causing spurious oscillations to occur in the spectral dimension. In the absence of such corrective measures as the present two methods, spectral ringing can manifest itself as systematic biases in some reconstructed spectral bands and can reduce the effectiveness of compression of spatially-low-pass subbands. One of the two methods is denoted mean subtraction. The basic idea of this method is to subtract mean values from spatial planes of spatially low-pass subbands prior to encoding, because (a) such spatial planes often have mean values that are far from zero and (b) zero-mean data are better suited for compression by methods that are effective for subbands of two-dimensional (2D) images. In this method, after the 3D wavelet decomposition is performed, mean values are computed for and subtracted from each spatial plane of each spatially-low-pass subband. The resulting data are converted to sign-magnitude form and compressed in a manner similar to that of a baseline hyperspectral- image-compression method. The mean values are encoded in the compressed bit stream and added back to the data at the appropriate decompression step. The overhead incurred by encoding the mean values only a few bits per spectral band is negligible with respect to the huge size of a typical hyperspectral data set. The other method is denoted modified decomposition. This method is so named because it involves a modified version of a commonly used multiresolution wavelet decomposition, known in the art as the 3D Mallat decomposition, in which (a) the first of multiple stages of a 3D wavelet transform is applied to the entire dataset and (b) subsequent stages are applied only to the horizontally-, vertically-, and spectrally-low-pass subband from the preceding stage. In the modified decomposition, in stages after the first, not only is the spatially-low-pass, spectrally-low-pass subband further decomposed, but also spatially-low-pass, spectrally-high-pass subbands are further decomposed spatially. Either method can be used alone to improve the quality of a reconstructed image (see figure). Alternatively, the two methods can be combined by first performing modified decomposition, then subtracting the mean values from spatial planes of spatially-low-pass subbands.