Chair: Arye Nehorai, Yale University (USA)
James J. Simpson, University of California-San Diego (USA)
The accurate analysis of complex satellite scenes is a critical component of many environmental studies. Unfortunately, satellite data often contain noise of various kinds which can compromise scientific analysis. Moreover, a satellite scene generally contains information on many different space scales associated with a variety of geophysical and/or biogeochemical processes. Thus, an accurate segmentation of such scenes is an essential step prior to scientific analysis. Only after such steps (e.g., noise reduction, segmentation) have been done can meaningful geophysical analyses be performed. This paper shows the natural synogestic relationship between the engineering and scientific components of the aforementioned problems. The benefits obtained by such a combined approach are illustrated with specific examples (oceanic, atmospheric, and terrestrial) using both polar orbiting and geostationary satellite data.
John C. Curlander, Vexcel Corporation (USA)
This paper describes a new technique for topographic mapping using an interferometric synthetic aperture radar (IFSAR). The IFSAR utilizes microwave interference for precise measurement of small linear displacements related to the surface topography. It is similar to a traditional SAR system except that it has two receive antennas (with essentially the same boresight) and the single recording channel is replaced by two independent channels, one for each antenna. From each imaging pass, the IFSAR measures (in addition to the backscattered power) the relative range distance from each antenna to the target pixel. The direct measurement is a relative phase difference which is related to the range diversity between the two antennas and the target area by the wavenumber. Using the relative phase difference in conjunction with the ranging and Doppler information, we can solve directly for the 3-D target location to produce a 3-D map of the target area without any earth model assumption.
John D. Gorman, Environmental Research Institute of Michigan (USA)
Nikola S. Subotic, Environmental Research Institute of Michigan (USA)
Brian J. Thelen, Environmental Research Institute of Michigan (USA)
We review the characteristics of hyperspectral imaging sensors and describe several important data exploitation applications in remote sensing. We then focus on a particular signal processing application, material identification, and propose a novel algorithm based on multiresolution wavelet techniques. Finally, we demonstrate the multiresolution material identification algorithm on data collected with a 211-band hyperspectral sensor.
James Ward, MIT Lincoln Laboratory (USA)
Advanced airborne radar systems are required to detect targets in the presence of both clutter and jamming. Ground clutter is extended in both angle and range, and is spread in Doppler frequency because of the platform motion. Space-time adaptive processing (STAP) refers to the simultaneous processing of the signals from an array antenna during a multiple pulse coherent waveform. STAP can provide improved detection of targets obscured by mainlobe clutter, detection of targets obscured by sidelobe clutter,and detection in combined clutter and jamming environments. Fully adaptive STAP is impractical for reasons of computational complexity and estimation with limited data, so partially adaptive approaches are required. This paper presents a taxonomy of partially adaptive STAP approaches that are classified according to the type of preprocessor, or equivalently, by the domain in which adaptive weighting occurs. Analysis of the rank of the clutter covariance matrix in each domain provides insight and conditions for preprocessor design.
John E. Molyneux, Widener University (USA)
We present a survey of theory and practice in tomographic imaging of subsurface features utilizing ground penetrating radar (GPR). The discussion will include: 1) A brief review of the equations governing radar scattering culminating in the Lippmann Schwinger (LS) equation; 2) a summary of approximations often made in developing GPR imaging algorithms based on inversion of the LS equation: the scalar wave approximation, the two dimensional scattering model, neglect of multiple scattering; 3) a discussion of certain inversion algorithms used in GPR work, including a short discussion of ill-posed problems utilizing a one dimensional model equation, a review of difficulties characteristic of tomographic inversion, a summary of results based on linearized inversion methods such as Devaney's generalized projection slice theorem, and finally, a brief overview of some numerical algorithms; and 4) a discussion of a recent field study using GPR to image shallow targets.
Gary Mavko, Stanford University (USA)
Nathalie Lucet, Stanford University (USA)
Tapan Mukerji, Stanford University (USA)
We present an integrated study in which we develop a range of techniques for deriving images of rock properties, such as porosity and shaliness, from cross-well seismic tomograms. One of the keys is to incorporate rock physics knowledge of the relations between velocity, porosity, and clay content which were developed in the laboratory and calibrated in the field. Geostatistical techniques, such as kriging and cokriging, are used as a means to combine the heterogeneous data set, consisting of well logs, the tomogram, and laboratory results.
David J. Thomson, AT&T Bell Laboratories (USA)
Possible changes in the world's climate resulting from human use of fossil fuel is perhaps the most serious problem facing the human race. The statistical problems in this area are also challenging; are we changing the climate in a measurable way, or are the currently perceived changes simply a result of natural variability? This paper outlines some of the statistical problems and a few examples that can be found in climate statistics.
Matthias G. Imhof, Massachusetts Institute of Technology (USA)
M. Nafi Tokso, Massachusetts Institute of Technology (USA)
The traditional approach in seismic imaging assumes homogeneous layers with constant seismic velocities. In real earth, layers have numerous small-scale variations of properties (e.g. the seismic P-wave velocity) which are so irregularly distribute that they can no longer be described by deterministic models; statistical approaches have to be used instead. They can be described only by their stochastic properties. In this paper, every distinctive layer is described by its mean velocity, its variance, and a spatial autocorrelation function. This velocity model is transformed into a model for the autocorrelation of the corresponding impulse response. The parameters for this model are then estimated by the minimum least square method using stacked and migrated seismic sections as input. The method is applied to a real seismic data set. Additional borehole data is used to estimate one of the parameters independently and thus to test the approach.