The mean-shift approach uses a local estimation of the pdf and moves every data points toward the modes. The direction is calculated from the mean value of surrounding points weighted by a Gaussian kernel. An advantage of the technique is that both radiometric and spatial information could be used in the weighted mean calculation. For polarimetric SAR images, we use likelihood ratio as radiometric similarity or distance measure. The spatial distance between pixels is also used with a Gaussian weight. Contours are well preserved because pixels on one side are dissimilar to pixels on the other side. To improve contour preservation, we examine how the tensor of pixel position can be integrated into the weight calculation. The tensor is calculated from weighted pixel position inside a window. Good PolSAR image smoothing is obtained.

Image segmentation and unsupervised classification are difficult problems. We propose to combine both. A clustering process is applied over segment mean values. Only large segments are considered. The clustering is composed of a mean-shift step and a hierarchical clustering step. The hierarchical grouping is based upon a powerful segmentation technique previously developed. The approach is applied on a 9-look polarimetric SAR image. Textured and non-textured image regions are considered. The K and Wishart distributions are used respectively. The unsupervised classification results can be very useful for image analysis and further supervised classification. The obtained region groups constitute an important simplification of the image.

In this paper, heterogeneous clutter models are introduced to describe Polarimetric Synthetic Aperture Radar (PolSAR) data. Based on the Spherically Invariant Random Vectors (SIRV) estimation scheme, the scalar texture parameter and the normalized covariance matrix are extracted. If the texture parameter is modeled by a Fisher PDF, the observed target scattering vector follows a KummerU PDF. Then, this PDF is implemented in a hierarchical segmentation algorithm. Segmentation results are shown on high resolution PolSAR data at L and X band.

Image segmentation and unsupervised classification are difficult problems. We propose to combine both. A clustering process is applied over segment mean values. Only large segments are considered. The clustering is composed of a mean-shift step and a hierarchical clustering step. The approach is applied on a 9-look polarimetric SAR image. Textured and non-textured image regions are considered. The K and Wishart distributions are used respectively. The obtained region groups constitute an important simplification of the image and a good initial classification map. Multiplying the class map by the image of scalar texture component produces an image almost identical to the original where speckle 'color' noise variation is filtered out.

The Polarimetric Synthetic Aperture Radar (PolSAR) covariance matrix is generally modeled by a complex Wishart distribution. For textured scenes, the product model is used and the texture component is often modeled by a Gamma distribution. In this paper, authors propose to use the Fisher distribution for texture modeling. From a Fisher distributed texture component, we derive the distribution of the complex covariance matrix and we propose to implement the KummerU distribution in a hierarchical segmentation and a hierarchical clustering algorithm. Segmentation and classification results are shown on synthetic images and on ESAR L-band PolSAR data over the Oberpfaffenhofen test-site.

The multilook polarimetric synthetic aperture radar (PolSAR) covariance matrix is generally modeled by a complex Wishart distribution. For textured areas, the product model is used, and the texture component is modeled by a Gamma distribution. In many cases, the assumption of Gamma-distributed texture is not appropriate. The Fisher distribution does not have this limitation and can represent a large set of texture distributions. As an example, we examine its advantage for an urban area. From a Fisher-distributed texture component, we derive the distribution of the complex covariance matrix for multilook PolSAR data. The obtained distribution is expressed in terms of the KummerU confluent hypergeometric function of the second kind. Those distributions are related to the Mellin transform and second-kind statistics (Log-statistics). The new KummerU-based distribution should provide in many cases a better representation of textured areas than the classicKdistribution. Finally, we show that the new model can discriminate regions with different texture distribution in a segmentation experiment with synthetic textured PolSAR images.

A hierarchical stepwise optimization process is developed for polarimetric synthetic aperture radar image segmentation. We show that image segmentation can be viewed as a likelihood approximation problem. The likelihood segment merging criteria are derived using the multivariate complex Gaussian, the Wishart distribution, and the K-distribution. In the presence of spatial texture, the Gaussian-Wishart segmentation is not appropriate. The K-distribution segmentation is more effective in textured forested areas. The validity of the product model is also assessed, and a field-adaptable segmentation strategy combining different criteria is examined.

The methods currently used for classification or segmentation of polarimetric SAR images are based on the multivariate complex Gaussian model. This should limit the application of these methods to "homogeneous" Gaussian areas, since their performances are significantly degraded in the presence of spatial texture. We show that image segmentation can be viewed as a likelihood approximation problem. The optimum criterion is derived for segmentation of K-distributed textured polarimetric SAR images. The product model is assessed and applied only within areas in which the model is valid. The new method is validated for ice type segmentation using Convair-580 SAR data collected in 1993 over Cornwallis Island in Canada.

A hierarchical stepwise optimization process is used for polarimetric SAR image segmentation. The process starts with small sets of pixels as segments and then sequentially merges the segment pair that minimises a stepwise criterion. The polarimetric information could be represented by a covariance matrix. The proposed criterion is based upon the testing of the equality of covariance matrices of adjacent regions. The segmentation of SAR images is greatly complicated by the presence of coherent speckle. We are using spatial constraints and contour shapes in order to improve the segmentation results.

A segmentation algorithm based on sequential optimization which produces a hierarchical decomposition of the picture is presented. The decomposition is data driven with no restriction on segment shapes. It can be viewed as a tree, where the nodes correspond to picture segments and where links between nodes indicate set inclusions. Picture segmentation is first regarded as a problem of piecewise picture approximation, which consists of finding the partition with the minimum approximation error. Then, picture segmentation is presented as an hypothesis-testing process which merges only segments that belong to the same region. A hierarchical decomposition constraint is used in both cases, which results in the same stepwise optimization algorithm. At each iteration, the two most similar segments are merged by optimizing a stepwise criterion. The algorithm is used to segment a remote-sensing picture, and illustrate the hierarchical structure of the picture.>