Leslie M. Novak

Optimal speckle reduction in polarimetric SAR imagery

Speckle is a major cause of degradation in synthetic aperture radar (SAR) imagery. With the availability of fully polarimetric SAR data, it is possible to use the three complex elements (HH, HV, VV) of the polarimetric scattering matrix to reduce speckle. The optimal method for combining the elements of the scattering matrix to minimize image speckle is derived, and the solution is shown to be a polarimetric whitening filter (PWF). A simulation of spatially correlated, K-distributed, fully polarimetric clutter is then used to compare the PWF with other, suboptimal speckle-reduction methods. Target detection performance of the PWF, span, and single-channel mod HH mod /sup 2/ detectors is compared with that of the optimal polarimetric detector (OPD). A novel, constant-false-alarm-rate (CFAR) detector (the adaptive PWF) is as a simple alternative to the OPD for detecting targets in clutter. This algorithm estimates the polarization covariance of the clutter, uses the covariance to construct the minimum-speckle image, and then tests for the presence of a target. An exact theoretical analysis of the adaptive PWF is presented; the algorithm is shown to have detection performance comparable with that of the OPD. >

Studies of target detection algorithms that use polarimetric radar data

Algorithms are described which make use of polarimetric radar information in the detection and discrimination of targets in a ground clutter background. The optimal polarimetric detector (OPD) is derived. This algorithm processes the complete polarization scattering matrix (PSM) and provides the best possible detection performance from polarimetric radar data. Also derived is the best linear polarimetric detector, the polarimetric matched filter (PMF), and the structure of this detector is related to simple polarimetric target types. New polarimetric target and clutter models are described and used to predict the performance of the OPD and the PME. The performance of these algorithms is compared with that of simpler detectors that use only amplitude information to detect targets. The ability to discriminate between target types by exploring differences in polarimetric properties is discussed. >

Identification of terrain cover using the optimum polarimetric classifier

A systematic approach for the identification of terrain media such as vegetation canopy, forest, and snow-covered fields is developed using the optimum polarimetric classifier. The covariance matrices for various terrain cover are computed from theoretical models of random medium by evaluating the scattering matrix elements. The optimal classification scheme makes use of a quadratic distance measure and is applied to classify a vegetation canopy consisting of both trees and grass. Experimentally measured data are used to validate the classification scheme. Analytical and Monte Carlo simulated classification errors using the fully polarimetric feature vector are compared with classification based on single features which include the phase difference between the VV and HH polarization returns. It is shown that the full polarimetric results are optimal and provide better classification performance than single feature measurements.

K-Distribution and Polarimetric Terrain Radar Clutter

A multivariate K- distribution is proposed to model the statistics of fully polarimetric radar data from earth terrain with polarizations HH, HV, VH, and VV. In this approach, correlated polarizations of radar signals, as characterized by a covariance matrix, are treated as the sum of N n- dimensional random vectors; N obeys the negative binomial distribution with a parameter α and mean N. Subsequently, an n- dimensional K- distribution, with either zero or nonzero mean, is developed in the limit of infinite N or illuminated area. The probability density function (PDF) of the K- distributed vector normalized by its Euclidean norm is independent of the parameter α and is the same as that derived from a zero-mean Gaussian-distributed random vector. The above model is well supported by experimental data provided by MIT Lincoln Laboratory and the Jet Propulsion Laboratory in the form of polarimetric measurements.

Effects of polarization and resolution on SAR ATR

Lincoln Laboratory is investigating the detection and classification of stationary ground targets using high resolution, fully polarimetric, synthetic aperture radar (SAR) imagery. A study is summarized in which data collected by the Lincoln Laboratory 33 GHz SAR were used to perform a comprehensive comparison of automatic target recognition (ATR) performance for several polarization/resolution combinations. The Lincoln Laboratory baseline ATR algorithm suite was used, and was optimized for each polarization/resolution case. Both the HH polarization alone and the optimal combination of HH, HV, and VV were evaluated; the resolutions evaluated were 1 ft/spl times/1 ft and 1 m/spl times/1 m. The data set used for this study contained approximately 74 km/sup 2/ of clutter (56 km/sup 2/ of mixed clutter plus 18 km/sup 2/ of highly cultural clutter) and 136 tactical target images (divided equally between tanks and howitzers).

Automatic target recognition using enhanced resolution SAR data

Using advanced technology, a new automatic target recognition (ATR) system has been developed that provides significantly improved target recognition performance compared with ATR systems that use conventional synthetic aperture radar (SAR) image-processing techniques. This significant improvement in target recognition performance is achieved by using a new superresolution image-processing technique that enhances SAR image resolution (and image quality) prior to performing target recognition. A computationally efficient two-level implementation of a template-based classifier is used to perform target recognition. The improvement in target recognition performance achieved using superresolution image processing in this new ATR system is quantified.

Fractional Brownian motion models for synthetic aperture radar imagery scene segmentation

The application of fractal random process models and their related scaling parameters as features in the analysis and segmentation of clutter in high-resolution, polarimetric synthetic aperture radar (SAR) imagery is demonstrated. Specifically, the fractal dimension of natural clutter sources, such as grass and trees, is computed and used as a texture feature for a Bayesian classifier. The SAR shadows are segmented in a separate manner using the original backscatter power as a discriminant. The proposed segmentation process yields a three-class segmentation map for the scenes considered in this study (with three clutter types: shadows, trees, and grass). The difficulty of computing texture metrics in high-speckle SAR imagery is addressed. In particular, a two-step preprocessing approach consisting of polarimetric minimum speckle filtering followed by noncoherent spatial averaging is used. The relevance of the resulting segmentation maps to constant-false-alarm-rate (CFAR) radar target detection techniques is discussed. >

Performance of 10- and 20-target MSE classifiers

MIT Lincoln Laboratory is responsible for developing the ATR (automatic target recognition) system for the DARPA-sponsored SAIP program; the baseline ATR system recognizes 10 GOB (ground order of battle) targets; the enhanced version of SAIP requires the ATR system to recognize 20 GOB targets. This paper presents ATR performance results for 10- and 20-target mean square error (MSE) classifiers using high-resolution SAR (synthetic aperture radar) imagery.

A multiresolution approach to discrimination in SAR imagery

We develop and test a new algorithm for discriminating man-made objects from natural clutter in synthetic-aperture radar (SAR) imagery. This algorithm exploits characteristic variations in speckle pattern as image resolution is varied from course to fine. We model these variations as an autoregression in scale, and then use the autoregressive model to define a multiresolution log-likelihood ratio discriminant. We incorporate this discriminant into the existing Lincoln Laboratory SAR system for automatic target recognition (ATR), and test the augmented system by applying it to millimeter-wave SAR imagery having 0.3 m resolution and representing 56 square kilometers of terrain. At a probability of detection of 0.95, the addition of the multiresolution discriminant reduces the number of natural-clutter false alarms by a factor of six.

Wavelet Based SAR Speckle Reduction and Image Compression

This paper evaluates the performance of the recently published wavelet-based algorithm for speckle reduction of SAR images. The original algorithm, based on the theory of wavelet thresholding due to Donoho and Johnstone, has been shown to improve speckle statistics. In this paper, we give more extensive results based on tests performed at Lincoln Laboratory (LL). The LL benchmarks show that the SAR imagery is significantly enhanced perceptually. Although the wavelet processed data results in an increase in the number of natural clutter false alarms, an appropriately modified CFAR detector (i.e., by clamping the estimated clutter standard deviation) eliminates the extra false alarms. The paper also gives preliminary results on the performance of the new and improved wavelet denoising algorithm based on the shift invariant wavelet transform. By thresholding the shift invariant discrete wavelet transform we can further reduce speckle to achieve a perceptually superior SAR image with ground truth information significantly enhanced. Preliminary results on the speckle statistics of this new algorithm is improved over the classical wavelet denoising algorithm. Finally, we show that the classical denoising algorithm as proposed by Donoho and Johnstone and applied to SAR has the added benefit of achieving about 3:1 compression with essentially no loss in image fidelity.