Search

This paper begins with some brief historical comments to set the stage for a discussion of the long-term potential for land remote sensing technology. This is followed by comments about what is needed to accelerate the achievement of this potential. The discussion is concluded with what concomitant development is needed with regard to a hyperspectral data analysis system. [ABSTRACT FROM AUTHOR]

In this paper, a new nonparametric feature extraction method is proposed for high-dimensional multiclass pattern recognition problems. It Is based on a nonparametric extension of scatter matrices. There are at least two advantages to using the proposed nonparametric scatter matrices. Pint, they are generally of full rank. This provides the ability to specify the number of extracted features desired and to reduce the effect of the singularity problem. This is in contrast to parametric discriminant analysis, which usually only can extract L - 1 (number of classes minus one) features. In a real situation, this may not be enough. Second, the nonparametric nature of scatter matrices reduces the effects of outliers and works well even for nonnormal datasets. The new method provides greater weight to samples near US expected decision boundary. This tends to provide for increased classification accuracy. [ABSTRACT FROM AUTHOR]

In this paper we proposes a supervised classifier based on implementation of the Bayes rule with kernels. The proposed technique first proposes an implicit nonlinear transformation of the data into a feature space seeking to fit normal distributions having a common covariance matrix onto the mapped data. One requirement of this approach is the evaluation of posterior probabilities. We express the discriminant function in dot-product form, and then apply the kernel concept to efficiently evaluate the posterior probabilities. The proposed technique gives the flexibility required to model complex data structures that originate from a wide range of class-conditional distributions. Although we end up with piecewise linear decision boundaries in the feature space, these corresponds to powerful nonlinear boundaries in the original input space. For the data we considered, we have obtained some encouraging results. [ABSTRACT FROM AUTHOR]

In this paper, we propose a cost-effective iterative semisupervised classifier based on a kernel concept. The proposed technique incorporates unlabeled data into the design of a binary classifier by introducing and optimizing a cost function in a feature space that maximizes the Rayleigh coefficient while minimizing the total cost associated with misclassified labeled samples. The cost assigned to misclassified labeled samples accounts for the number of misclassified labeled samples as well as the amount by which they are on the wrong side of the boundary, and this counterbalances any potential adverse effect of unlabeled data on the classifier performance. Several experiments performed with remotely sensed data demonstrate that using the proposed semisupervised classifier shows considerable improvements over the supervised-only counterpart. [ABSTRACT FROM AUTHOR]

It is well known that there is a strong relation between class definition precision and classification accuracy in pattern classification applications. In hyperspectral data analysis, usually classes of interest contain one or more components and may not be well represented by a single Gaussian density function. In this paper, a model-based mixture classifier, which uses mixture models to characterize class densities, is discussed. However, a key outstanding problem of this approach is how to choose the number of.components and determine their parameters for such models in practice, and to do so in the face of limited training sets where estimation error becomes a significant factor. The proposed classifier estimates the number of subclasses and class statistics simultaneously by choosing the best model. The structure of class covariances is also addressed through a model-based covariance estimation technique introduced in this paper. [ABSTRACT FROM AUTHOR]

MultiSpec is a multispectral image data analysis software application. It is intended to provide a fast, easy-to-use means for analysis of multispectral image data, such as that from the Landsat, SPOT, MODIS or IKONOS series of Earth observational satellites, hyperspectral data such as that from the Airborne Visible–Infrared Imaging Spectrometer (AVIRIS) and EO-1 Hyperion satellite system or the data that will be produced by the next generation of Earth observational sensors. The primary purpose for the system was to make new, otherwise complex analysis tools available to the general Earth science community. It has also found use in displaying and analyzing many other types of non-space related digital imagery, such as medical image data and in K-12 and university level educational activities.MultiSpec has been implemented for both the Apple Macintosh® and Microsoft Windows® operating systems (OS). The effort was first begun on the Macintosh OS in 1988. The GLOBE (http://www.globe.gov) program supported the development of a subset of MultiSpec for the Windows OS in 1995. Since then most (but not all) of the features in the Macintosh OS version have been ported to the Windows OS version.Although copyrighted, MultiSpec with its documentation is distributed without charge. The Macintosh and Windows versions and documentation on its use are available from the World Wide Web at URL:http://dynamo.ecn.purdue.edu/∼biehl/MultiSpec/MultiSpec is copyrighted (1991–2001) by Purdue Research Foundation, West Lafayette, Indiana 47907. [Copyright &y& Elsevier]

There are many factors to consider in carrying out a hyperspectral data classification. Perhaps chief among them are class training sample size, dimensionality, and distribution separability. The intent of this study is to design a classification procedure that is robust and maximally effective, but which provides the analyst with significant assists, thus simplifying the analyst's task. The result is a quadratic mixture classifier based on Mixed-LOOC2 regularized discriminant analysis and nonparametric weighted feature extraction. This procedure has the advantage of providing improved classification accuracy compared to typical previous methods but requires minimal need to consider the factors mentioned above. Experimental results demonstrating these Properties are presented. [ABSTRACT FROM AUTHOR]

In this paper, an adaptive Bayesian contextual classification procedure that utilizes both spectral and spatial interpixel dependency contexts in estimation of statistics and classification is proposed. Essentially, this classifier is the constructive coupling of an adaptive classification procedure and a Bayesian contextual classification procedure. In this classifier, the joint prior probabilities of the classes of each pixel and its spatial neighbors are modeled by the Markov random field. The estimation of statistics and classification are performed in a recursive manner to allow the establishment of the positive-feedback process in a computationally efficient manner. Experiments with real hyperspectral data show that, starting with a small training sample set, this classifier can reach classification accuracics similar to that obtained by a pixelwise maximum likelihood pixel classifier with a very large training sample set. Additionally, classification maps are produced that have significantly less speckle error. [ABSTRACT FROM AUTHOR]

Presents a study that proposed a family of adaptive covariance estimators to mitigate the problem of limited training samples for application to hyperspectral data analysis in quadratic maximum likelihood classification. Forms of the covariance estimation depending on the form of covariance matrices used; Discussion on how to produce a family of adaptive covariance estimators; Mean classification accuracy for each estimator.

Presents a study that developed an improved regularized covariance estimator of each class with the advantages of Leave-One-Out Covariance Estimator (LOOC) and Bayesian LOOC. Information on mixed leave-one-out covariance estimators; Ways to overcome estimation errors; Experiment design for comparing LOOC, mixed-LOOC1 and mixed-LOOC2.

Proposes a process methodology for remote sensing data analysis to enhance understanding of the terrestrial surface. Function of a mathematical modeling on the computer; Principles of the proposed methodology; Phases of a remote sensing data analysis project; Significance of the proposed methodology.

Presents a study which proposed a self-learning and self-improving adaptive classifier. Potential benefits of the adaptive classifier; Estimation of the parameters of a mixture of two normal distributions; Bound on the probability of error.

Presents a study which proposed a robust parameter estimation method for remote sensing problems. Expectation maximization algorithm; Robust estimation; Experimental results; Conclusion.

Presents information on a study of hyperspectral data analysis using the projection pursuit method. Projection pursuit and dimension reduction; Preprocessing block stages and the initial conditions; Conclusion.

Presents information on a study which addressed a classification problem in which class definition through training samples is provided a priori only for a particular class of interest using weighted unsupervised clustering. Partially supervised classification using unsupervised clustering with weights; Experiments and discussion; Conclusions.

Presents information on a study which deals with some aspects of the problem of classifying remote sensing digital image data in which classes with equal or nearly equal first-order statistics are present. Component of covariance matrix in multispectral image classification; Design of a statistically based classifier; Geometrical behavior of the second-order statistics in remote sensing image data classification; Conclusions.

Presents information on a study which described a covariance estimator formulated under an empirical Bayesian setting to mitigate the problem of limited training samples in the Gaussian maximum likelihood classification of remote sensing. Information on the covariance estimator based on a Bayesian formulation; Experimental results; Discussion and conclusions.

Proposes two decision fusion-based multitemporal classifiers, the jointly likelihood and the weighted majority fusion. Discussion of the multitemporal classification; Decision fusion in multitemporal classification; Description of the multitemporal data sets and training.

Develops a new covariance matrix estimator useful for designing classifiers with limited training data. High dimensional data; Gaussian maximum likelihood classification; Definition of the covariance matrix estimator.

A foreword to "Journal of Terrestrial Observation" is presented.

Introduces a series of articles that aims to describe advances in processing and analysis techniques for hyperspectral data as of July 2001. Dense sampling of the spectral range of the electromagnetic spectrum; Suboptimal methods for feature extraction; Spatial resolution of satellite hyperspectral images.