&lt;title&gt;Application of stochastic mixing models to hyperspectral detection problems&lt;/title&gt;

Stocker, Alan D.; Schaum, A.

doi:10.1117/12.280584

Cited by 64 publications

(57 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…1 and 2 suggest that improved anomaly detection could result from using a bank of subspace detectors rather than a global anomaly detector, and methods of defining this collection of subspaces should be explored. Further improvements in anomaly detection may result from using improved clutter models as suggested by the work on stochastic mixture models that unifies the linear mixture and Gaussian mixture ap- proaches [34], [35]. Furthermore, the requirements for image registration in the application of Chronochrome need to be established.…”

Section: Discussionmentioning

confidence: 99%

“…Since many of the conditions and parameters that govern performance are unknown or poorly characterized in an operational setting, it can be very difficult to dynamically select the optimal detection algorithm. To obtain more consistent anomaly detection performance from a single algorithm, one might investigate more general statistical models [34], [35]. Alternatively, we construct a multiple-algorithm-decision or fusion criterion that combines multiple discrimination features.…”

Section: Multiple Algorithm Fusionmentioning

confidence: 99%

See 1 more Smart Citation

Anomaly detection from hyperspectral imagery

Stein

Beaven

Hoff³

et al. 2002

IEEE Signal Process. Mag.

631

260

View full text Add to dashboard Cite

H yperspectral sensors are a new class of optical sensor that collect a spectrum from each point in a scene. They differ from multispectral sensors in that the number of bands is much higher (20 or more) and the spectral bands are contiguous. For remote sensing applications, they are typically deployed on either aircraft or satellites. The data product from a hyperspectral sensor is a three-dimensional array or "cube" of data with the width and length of the array corresponding to spatial dimensions and the spectrum of each point as the third dimension. While this data cube is a convenient way to view the product, at most two of the dimensions can be acquired simultaneously. In the most common configuration, the spectral and one spatial dimension are acquired simultaneously to give a high quality spectral signature for each point in the scene. AVIRIS [1] and many other sensors directed toward terrain classification are flying spot sensors that acquire an image by raster scanning the scene in width while the platform moves to build up the second spatial dimension. More recently, the spatial resolution of hyperspectral sensors has been improved by using a two-dimensional focal plane array that simultaneously acquires a spectrum along a line of points in the scene. The second dimension is then built up by motion of the platform.

show abstract

Section: Discussionmentioning

confidence: 99%

Section: Multiple Algorithm Fusionmentioning

confidence: 99%

Anomaly detection from hyperspectral imagery

Stein

Beaven

Hoff³

et al. 2002

IEEE Signal Process. Mag.

631

260

View full text Add to dashboard Cite

show abstract

“…As a consequence, the proposed methodology has to be coupled with one of the many identification techniques to estimate endmember spectra. These techniques include geometrical methods [6,7] or statistical procedures [8,9].…”

Section: Linear Mixing Modelmentioning

confidence: 99%

Spectral Unmixing of Hyperspectral Images using a Hierarchical Bayesian Model

Dobigeon

Tourneret

2007

2007 IEEE International Conference on Acoustics, Speech and Signal Processing - ICASSP '07

View full text Add to dashboard Cite

This paper addresses the problem of hyperspectral image unmixing. A new hierarchical Bayesian algorithm is proposed to estimate the coefficients of a linear mixture of spectra associated to a given pixel of the image. Appropriate priors are introduced to guaranty the positivity and additivity constraints inherent to the mixture coefficients. These coefficients referred to as abundances are then estimated from their posterior following the principles of Bayesian inference. The estimation is performed by using a Gibbs sampling strategy which generates samples distributed according the abundance posterior distribution. These samples are then averaged yielding the abundance minimum mean square error estimator.Index Terms-Bayesian inference, Monte Carlo methods, spectral unmixing, hyperspectral images.

show abstract

“…Clearly, g(* +1 ) consists of an adjustment to g^ based on the weighted sum of the parameter vector error,(g( fc ) -g' 0 '), and the spectrum error, (x -S^x.^). In the case of both errors, the difference is measured with reference to the starting values in g(°), not the previous value of g. The implication here is that dramatic changes in estimates, when gauged against the values in g, are inversely weighted by the associated confidence in the original estimates appearing in r s (o)"(o) • Gaussian Class Estimation An adaptation of the linear mixing model combines the geometric approach in Section 3.3.1.1 with stochastic data clustering approaches [30]. The stochastic mixing model (SMM) introduces the concept of hard endmember classes as the stochastic extension of deterministic endmembers and assumes that all data in a scene is Gaussian and arises from a linear combination of at least one hard class.…”

Section: Maximum Likelihood Estimationmentioning

confidence: 99%

A Taxonomy of Spectral Unmixing Algorithms and Performance Comparisons

Keshava¹

2002

View full text Add to dashboard Cite

In this report, algorithms for spectral unmixing are organized into taxonomies and their performance is then compared. Our motivation is to collectively organize and relate algorithms in order to assess the current state-of-the-art in the field and to facilitate objective comparisons between methods. The hyperspectral sensing community is populated by investigators with disparate scientific backgrounds and efforts in spectral unmixing developed within disparate communities have inevitably led to duplication. This report is intended to remove ambiguity and redundancy by using a standard vocabulary, and clearly summarize what has and has not been done. As will be evident, the framework for the taxonomies derives its organization from the fundamental, philosophical assumptions imposed on the problem, rather than the common calculations they perform, or the similar outputs they might yield. The taxonomies are supplemented by a comparison of unmixing performance using techniques that typify the approaches of wide classes of algorithms.

show abstract

<title>Application of stochastic mixing models to hyperspectral detection problems</title>

Cited by 64 publications

References 0 publications

Anomaly detection from hyperspectral imagery

Anomaly detection from hyperspectral imagery

Spectral Unmixing of Hyperspectral Images using a Hierarchical Bayesian Model

A Taxonomy of Spectral Unmixing Algorithms and Performance Comparisons

Contact Info

Product

Resources

About