Mitchell R. Grunes

Unsupervised classification using polarimetric decomposition and the complex Wishart classifier

The authors propose a new method for unsupervised classification of terrain types and man-made objects using polarimetric synthetic aperture radar (SAR) data. This technique is a combination of the unsupervised classification based on polarimetric target decomposition, S.R. Cloude et al. (1997), and the maximum likelihood classifier based on the complex Wishart distribution for the polarimetric covariance matrix, J.S. Lee et al. (1994). The authors use Cloude and Pottiers method to initially classify the polarimetric SAR image. The initial classification map defines training sets for classification based on the Wishart distribution. The classified results are then used to define training sets for the next iteration. Significant improvement has been observed in iteration. The iteration ends when the number of pixels switching classes becomes smaller than a predetermined number or when other criteria are met. The authors observed that the class centers in the entropy-alpha plane are shifted by each iteration. The final class centers in the entropy-alpha plane are useful for class identification by the scattering mechanism associated with each zone. The advantages of this method are the automated classification, and the interpretation of each class based on scattering mechanism. The effectiveness of this algorithm is demonstrated using a JPL/AIRSAR polarimetric SAR image.

Polarimetric SAR speckle filtering and its implication for classification

This paper proposes a new approach in polarimetric synthetic aperture radar (SAR) speckle filtering. The new approach emphasizes preserving polarimetric properties and statistical correlation between channels, not introducing crosstalk, and not degrading the image quality. In the last decade, speckle reduction of polarimetric SAR imagery has been studied using several different approaches. All of these approaches exploited the degree of statistical independence between linear polarization channels. The preservation of polarimetric properties and statistical characteristics such as correlation between channels were not carefully addressed. To avoid crosstalk, each element of the covariance matrix must be filtered independently. This rules out current methods of polarimetric SAR filtering. To preserve the polarimetric signature, each element of the covariance matrix should be filtered in a way similar to multilook processing by averaging the covariance matrix of neighboring pixels. However, this must be done without the deficiency of smearing the edges, which degrades image quality and corrupts polarimetric properties. The proposed polarimetric SAR filter uses edge-aligned nonsquare windows and applies the local statistics filter. The impact of using this polarimetric speckle filtering on terrain classification is quite dramatic in boosting classification performance. Airborne polarimetric radar images are used for illustration.

Classification of multi-look polarimetric SAR imagery based on complex Wishart distribution

Abstract Multi-look polarimetric SAR (synthetic aperture radar) data can be represented either in Mueller matrix form or in complex covariance matrix form. The latter has a complex Wishart distribution. A maximum likelihood classifier to segment polarimetric SAR data according to terrain types has been developed based on the Wishart distribution. This algorithm can also be applied to multifrequency multi-look polarimetric SAR data, as well as 10 SAR data containing only intensity information. A procedure is then developed for unsupervised classification. The classification error is assessed by using Monte Carlo simulation of multilook polarimetric SAR data, owing to the lack of ground truth for each pixel. Comparisons of classification errors using the training sets and single-look data are also made. Applications of this algorithm are demonstrated with NASA/JPL P-, L- and C-band polarimetric SAR data.

Quantitative comparison of classification capability: fully polarimetric versus dual and single-polarization SAR

This paper addresses the land-use classification capabilities of fully polarimetric synthetic aperture radar (SAR) versus dual-polarization and single-polarization SAR for P-, L-, and C-Band frequencies. A variety of polarization combinations will be investigated for application to crop and tree age classification. Based on the complex Wishart distribution for the covariance matrix, maximum likelihood (ML) classifiers for all polarization combinations were used to assess quantitative classification accuracy. Thus, this allows optimally selecting the frequency and the combination of polarizations for various applications.

A new technique for noise filtering of SAR interferometric phase images

This paper addresses the noise filtering problem for SAR interferogram phase images. The phase noise is characterized by an additive noise model, and a filtering algorithm based on this noise model was developed by filtering noise along fringes. In addition, this filter adaptively adjusts the amount of filtering according to the coherence. The effectiveness of this filter is demonstrated using SIR-C/X-SAR multi-pass generated interferograms.

Speckle reduction in multipolarization, multifrequency SAR imagery

An algorithm to take advantage of this polarization diversity to suppress the speckle effect with much less resolution broadening than using spatial filtering is discussed. The coupling between polarization channels is minimized by using local intensity ratios. The degree of speckle reduction is similar to two-look or three-look processing. The same algorithm can also be used to process multifrequency polarimetric SAR. For three-frequency aircraft SAR data speckle reduction equivalent to six-look processing can be achieved. Further speckle reduction is possible by applying speckle filters in the spatial domain. In addition, a vector speckle filter which operates simultaneously in the polarization and spatial domains is tested. Experimental results with simulated polarimetric SAR as well as one-look and multilook parametric SAR data demonstrate the effectiveness of these speckle reductions, with minimum resolution broadening and coupling between polarimetric and frequency channels. Comparisons with other algorithms are also made. >

Scattering-model-based speckle filtering of polarimetric SAR data

A new concept in polarimetric synthetic aperture radar (POLSAR) speckle filtering that preserves the dominant scattering mechanism of each pixel is proposed in this paper. The basic principle is to select pixels of the same scattering characteristics to be included in the filtering process. To achieve this, the algorithm first applies the Freeman and Durden decomposition to separate pixels into three dominant scattering categories: surface, double bounce, and volume, and then unsupervised classification is applied. Speckle filtering is performed using the classification map as a mask. A single-look or multilook pixel centered in a 9 /spl times/ 9 window is filtered by including only pixels in the same and two neighboring classes from the same scattering category. This filter is effective in speckle reduction, while perfectly preserving strong point target signatures, and retains edges, linear, and curved features in the POLSAR data. The effect of speckle filtering on scattering characteristics, such as entropy, anisotropy, and alpha angle, will be discussed.

Unsupervised classification using polarimetric decomposition and complex Wishart classifier

The authors propose a new method for unsupervised classification of terrain types and man-made objects using polarimetric SAR data. This technique is a combination of the unsupervised classification based on the polarimetric target decomposition (Cloude and Pottier, 1997) and the maximum likelihood classifier based on the complex Wishart distribution (Lee et al., 1994). The advantage of this approach is that clusters may be identified by the scattering mechanisms from the target decomposition. The effectiveness of this algorithm is demonstrated using JPL/AIRSAR and SIR-C polarimetric SAR images.

Speckle filtering and coherence estimation of polarimetric SAR interferometry data for forest applications

Recently, polarimetric synthetic aperture radar (SAR) interferometry has generated much interest for forest applications. Forest heights and ground topography can be extracted based on interferometric coherence using a random volume over ground coherent mixture model. The coherence estimation is of paramount importance for the accuracy of forest height estimation. The coherence (or correlation coefficient) is a statistical average of neighboring pixels of similar scattering characteristics. The commonly used algorithm is the boxcar filter, which has the deficiency of indiscriminate averaging of neighboring pixels. The result is that coherence values are lower than they should be. In this paper, we propose a new algorithm to improve the accuracy in the coherence estimation based on speckle filtering of the 6/spl times/6 polarimetric interferometry matrix. Simulated images are used to verify the effectiveness of this adaptive algorithm. German Aerospace Center (DLR) L-Band E-SAR data are applied to demonstrate the improved accuracy in coherence and in forest height estimation.

Integration of optical and radar classifications for mapping pasture type in Western Australia

In this study, independent classifications of Landsat Thematic Mapper imagery and Jet Propulsion Laboratory AirSAR were combined to create an integrated classification of pasture and other vegetation types for a study area in the agricultural zone of Western Australia. The resulting classification combines greenness and brightness information from optical data with structure and water content information from synthetic aperture radar (SAR). Field observations of vegetation type, botanical composition, ground cover percentage, wet and dry biomass, canopy height, and soil water content were collected at 34 sites representing a range of pastures, browse shrubs, and crops. An unsupervised version of the Complex Wishart classification procedure, based on preserving scattering characteristics from the Freeman and Durden backscatter decomposition, was applied to the C-, L-, and P-band polarimetric SAR data. The optical classification was carried out using a principle component analysis on the green, red, and near-infrared bands and clustering on the basis of a class centroid distance measure and knowledge of ground targets. These two classification results were then fused together. Assessment of a confusion matrix using the individual sites showed that identification of more uniform, dense, and structurally distinct canopies was better than that of more diverse, sparse, and structurally ambiguous canopies, as the former were better represented by the canopy height attribute used in the SAR classification component. The optical classification enabled correction of SAR misclassification of vegetation due to surface roughness and soil moisture effects, or similar backscatter responses from herbaceous or arboreal canopies. The results show that simplification of vegetation into groups based upon properties with sensitive responses in both the optical and SAR domains, and combination of separate SAR and optical classifications, has potential for improving classification of diverse and heterogeneous herbaceous and browse cover in grazing lands. However, collection of ground calibration data must be at an appropriate spatial scale and include canopy and surface measurements directly related to backscatter mechanisms and spectral sensitivity.