Multitemporal satellite synthetic aperture radar (SAR) images are a useful source of information for geophysicists to monitor changing regions. In this paper, a new approach is proposed to extract from multitemporal SAR images two kinds of information: temporal changes (flooded areas, coastline erosion, etc.) and stable spatial features (roads, rivers, etc.). The novelty of the proposed approach is to detect simultaneously these two kinds of discontinuities. In a first step, the contrast and the heterogeneity information is extracted by a "multitemporal" application of the ratio of local means and by new three-dimensional texture parameters based on the log-cumulants. In a second step, the resulting attributes that measure the time variability or the presence of spatial features are merged. An interactive fuzzy fusion approach is proposed to provide end-users with a simple and easily understandable tool for tuning the change-detection results. The performances of the proposed attributes and fusion technique are presented on a set of seven multitemporal SAR images acquired by the European Remote Sensing (ERS-1) satellite.

Application of log-cumulants to the detection of spatiotemporal discontinuities in multitemporal SAR images

In this paper, we develop a novel classification approach for multiresolution, multisensor [optical and synthetic aperture radar (SAR)], and/or multiband images. This challenging image processing problem is of great importance for various remote sensing monitoring applications and has been scarcely addressed so far. To deal with this classification problem, we propose a two-step explicit statistical model. We first design a model for the multivariate joint class-conditional statistics of the coregistered input images at each resolution by resorting to multivariate copulas. Such copulas combine the class-conditional marginal probability density functions (pdfs) of each input channel that are estimated by finite mixtures of well-chosen parametric families. We consider different distribution families for the most common types of remote sensing imagery acquired by optical and SAR sensors. We then plug the estimated joint pdfs into a hierarchical Markovian model based on a quad-tree structure, where each tree-scale corresponds to the different input image resolutions and to corresponding multiscale decimated wavelet transforms, thus preventing a strong resampling of the initial images. To obtain the classification map, we resort to an exact estimator of the marginal posterior mode. We integrate a prior update in this model in order to improve the robustness of the developed classifier against noise and speckle. The resulting classification performance is illustrated on several remote sensing multiresolution data sets, including very high resolution and multisensor images acquired by COSMO-SkyMed and GeoEye-1.

/pdf/supervised-classification-of-multisensor-and-multiresolution-4a0ksff7rl.pdf

Supervised Classification of Multisensor and Multiresolution Remote Sensing Images With a Hierarchical Copula-Based Approach

Change detection (CD) in heterogeneous multitemporal satellite images is an emerging and challenging topic in remote sensing. In particular, one of the main challenges is to tackle the problem in an unsupervised manner. In this paper, we propose an unsupervised framework for bitemporal heterogeneous CD based on the comparison of affinity matrices and image regression. First, our method quantifies the similarity of affinity matrices computed from colocated image patches in the two images. This is done to automatically identify pixels that are likely to be unchanged. With the identified pixels as pseudotraining data, we learn a transformation to map the first image to the domain of the other image and vice versa. Four regression methods are selected to carry out the transformation: Gaussian process regression, support vector regression, random forest regression (RFR), and a recently proposed kernel regression method called homogeneous pixel transformation. To evaluate the potentials and limitations of our framework and also the benefits and disadvantages of each regression method, we perform experiments on two real data sets. The results indicate that the comparison of the affinity matrices can already be considered a CD method by itself. However, image regression is shown to improve the results obtained by the previous step alone and produces accurate CD maps despite of the heterogeneity of the multitemporal input data. Notably, the RFR approach excels by achieving similar accuracy as the other methods, but with a significantly lower computational cost and with fast and robust tuning of hyperparameters.

/pdf/unsupervised-image-regression-for-heterogeneous-change-ftonoxtt13.pdf

Unsupervised Image Regression for Heterogeneous Change Detection

Change detection in heterogeneous multitemporal satellite images is an emerging and challenging topic in remote sensing. In particular, one of the main challenges is to tackle the problem in an unsupervised manner. In this paper we propose an unsupervised framework for bitemporal heterogeneous change detection based on the comparison of affinity matrices and image regression. First, our method quantifies the similarity of affinity matrices computed from co-located image patches in the two images. This is done to automatically identify pixels that are likely to be unchanged. With the identified pixels as pseudo-training data, we learn a transformation to map the first image to the domain of the other image, and vice versa. Four regression methods are selected to carry out the transformation: Gaussian process regression, support vector regression, random forest regression, and a recently proposed kernel regression method called homogeneous pixel transformation. To evaluate the potentials and limitations of our framework, and also the benefits and disadvantages of each regression method, we perform experiments on two real data sets. The results indicate that the comparison of the affinity matrices can already be considered a change detection method by itself. However, image regression is shown to improve the results obtained by the previous step alone and produces accurate change detection maps despite of the heterogeneity of the multitemporal input data. Notably, the random forest regression approach excels by achieving similar accuracy as the other methods, but with a significantly lower computational cost and with fast and robust tuning of hyperparameters.

/pdf/unsupervised-image-regression-for-heterogeneous-change-lshkiyqfse.pdf

Change detection in heterogeneous bitemporal satellite images has become an emerging, important, and challenging research topic in remote sensing for rapid damage assessment. In this article, we explore a new parametric mapping strategy based on a modified geometric fractal decomposition and a contractive mapping approach allowing us to project the before image on any after imaging modality type. This projection exploits the fact that any satellite image data can be approximatively encoded in terms of spatial self-similarities at different scales and this property remains quite invariant to a given imaging modality type. Once the projection is performed and that a pixelwise difference map between the two images (presented in the same imaging modality) is then binarized in the unsupervised Bayesian framework. At this stage, we will test several parameter estimation procedures combined with several segmentation strategies based on different Bayesian cost functions. The experiments for change detection, with real images showing different multimodalities and changed events, indicate that this new fractal-based projection method, which is entirely based on a series of structural and spatial information, is an interesting alternative to classical regression-based projection methods (based only on luminance transformation). Besides, the experiments also show that the difference map, resulting in this novel projection strategy, is also particularly amenable for an unsupervised Markovian binarization approach.

A Fractal Projection and Markovian Segmentation-Based Approach for Multimodal Change Detection

With the ever-increasing number and diversity of Earth observation satellites, it steadily becomes more important to be able to analyze compound data sets consisting of different types of images acquired by different sensors. In this paper, we examine different ways of obtaining joint distributions of such images, and we propose a method that enables incorporation of correlations between images while keeping a good fit to the marginal distributions. The approach basically consists of two steps. First, the marginal densities are specified. Based on this specification, each marginal variable is transformed to a normal distributed variable. The joint distribution of the transformed variables is assumed to be multivariate normal. Transforming back to the original scale gives a joint distribution with dependence, where the initial marginal distributions are preserved. The parameters of the new joint distribution can be estimated. The focus is on marginal distributions that are Gamma, K, or Gaussian, although any distribution could be considered. The joint distributions produced by the transformation method can be used in supervised classification of radar and optical images. Results obtained for a set of four-look synthetic aperture radar (SAR) images, as well as a combination of SAR and optical images, are presented.

On the Combination of Multisensor Data Using Meta-Gaussian Distributions

For a given ground cover class, there is no straightforward way of expressing the joint distribution of a set of correlated radar images represented in amplitude or intensity. In this article we propose a general transformation method that permits incorporation of inter-image covariance while keeping a good fit to the marginal distributions. The approach is here studied for Gamma marginals, and the results of tests on a multitemporal series of ERS-1 multi-look images are presented.

Joint distributions for correlated radar images

For a given ground cover class, there is no straight-forward way of expressing the joint distribution of a set of correlated radar images represented in amplitude or intensity. In this article we propose a general transformation method that permits incorporation of inter-image covariance while keeping a good fit to the marginal distributions. This meta-Gaussian approach is studied here for Gamma marginals, and the results of tests on a multi-temporal series of ERS-1 multi-look images are presented.

B. Storvik

Papers

On the Combination of Multisensor Data Using Meta-Gaussian Distributions

Joint distributions for correlated radar images

Joint Distributions for Multi-Temporal Series of Radar Images