Image synthesis tools provide a means for generating hyperspectral image data without the expense of data
collection. An important use of these tools is to provide data for the assessment of image exploitation algorithms.
However, the detailed spectral/spatial structure of synthetic images is typically not sufficiently realistic to support
the prediction of algorithm performance on real data. In this paper, we develop a new method for hyperspectral
texture synthesis that accurately simulates the spectral/spatial structure of real hyperspectral image data. We
demonstrate the utility of the new technique by presenting real and synthesized images and by analyzing spectral
angle deviation from the mean curves that describe spectral properties. We also demonstrate that a signaturebased
detection algorithm has similar performance against real and synthesized hyperspectral backgrounds.
A 3-D spectral/spatial DFT represents an image region using a dense sampling in the frequency domain. An alternative
approach is to represent a 3-D DFT by its projection onto a set of functions that capture specific orientation, scale, and
spectral attributes of the image data. For this purpose, we have developed a new model for spectral/spatial information
in images based on three-dimensional Gabor filters. This model achieves optimal joint localization in space and
frequency and provides an efficient means of sampling a three-dimensional frequency domain representation of HSI
data. Since 3-D Gabor filters allow for a large number of spectral/spatial quantities to be used to represent an image
region, the performance and efficiency of algorithms that use this representation can be improved if methods are
available to reduce the dimensionality of the model. Thus, we have derived methods for selecting filters that emphasize
the most significant spectral/spatial differences between the various classes in a scene. We demonstrate the utility of the
new model for region classification in AVIRIS data.
In this work, we develop a new method for multispectral and hyperspectral texture synthesis using the multiband
distribution and power spectral densities. Different approaches to this problem are mostly case specific and
include histogram explosion, equalization in HSV or some other color space, and equalization based on the
earth mover distance. For multiband images, the usual practice is to define the power spectral density for
each band separately. While this captures the in-band autocorrelations, the cross-band correlations are not
captured. Sometimes cross-psds are defined if it is known that cross-band correlations are important. However,
as the number of bands increases, this method becomes computationally prohibitive. We propose a method that
expresses psds for multiband images using a 3D fourier transform. An iterative scheme is used to equalize the
histogram and psds for an input and target image. Our experiments show that the iteration tends to converge
after 5-10 steps. The proposed method is computationally efficient and yields satisfactory results. We compare
synthesized multispectral textures with real multispectral data.
In this paper, we design a decision rule to select optimized neighbor sets for multispectral images. We assume that multispectral images can be modeled by parametric Gaussian Random Fields. From a class of such models with different neighbor sets, we choose the best representation employing bayesian methods. The chosen model accounts for interactions within each of the spectral bands as well as the interaction between different spectral bands in a multispectral image. We evaluate the performance of the neighbor sets for multispectral texture classification.
We present a generalized random field model in a random environment to classify hyperspectral textures. The model generalizes traditional random field models by allowing the spatial interaction parameters of the field to be random variables. Principal component analysis is used to reduce the dimensionality of the data set to a small number of spectral bands that capture almost all of the energy in the original hyperspectral textures. Using the model we obtain a compact feature vector that efficiently computes within- and between-band information. Using a set of hyperspectral samples, we evaluate the performance of this model for classifying textures and compare the results with other approaches that consider different kinds of spatial, spectral, and intensity distribution information.
We present a generalized random field model in a random environment to classify hyperspectral textures. The model generalizes traditional random field models by allowing the spatial interaction parameters of the field to be random variables. Principal component analysis is used to reduce the dimensionality of the data set to a small number of spectral bands that caputure almost all of the energy in the original hyperspectral textures. Using the model we obtain a compact feature vector that efficiently computes within and between band information. Using a set of hyperspectral samples, we evaluate the performance of this model for classifying textures and compare the results with other approaches.
Access to the requested content is limited to institutions that have purchased or subscribe to SPIE eBooks.
You are receiving this notice because your organization may not have SPIE eBooks access.*
*Shibboleth/Open Athens users─please
sign in
to access your institution's subscriptions.
To obtain this item, you may purchase the complete book in print or electronic format on
SPIE.org.
INSTITUTIONAL Select your institution to access the SPIE Digital Library.
PERSONAL Sign in with your SPIE account to access your personal subscriptions or to use specific features such as save to my library, sign up for alerts, save searches, etc.