Alejandro C. Frery

Information fusion for wireless sensor networks: Methods, models, and classifications

Wireless sensor networks produce a large amount of data that needs to be processed, delivered, and assessed according to the application objectives. The way these data are manipulated by the sensor nodes is a fundamental issue. Information fusion arises as a response to process data gathered by sensor nodes and benefits from their processing capability. By exploiting the synergy among the available data, information fusion techniques can reduce the amount of data traffic, filter noisy measurements, and make predictions and inferences about a monitored entity. In this work, we survey the current state-of-the-art of information fusion by presenting the known methods, algorithms, architectures, and models of information fusion, and discuss their applicability in the context of wireless sensor networks.

A model for extremely heterogeneous clutter

A new class of distributions, G distributions, arising from the multiplicative model is presented, along with their main properties and relations. Their densities are derived for complex and multilook intensity and amplitude data. Classical distributions, such as K, are particular cases of this new class. A special case of this class called G/sup 0/, that has as many parameters as K distributions, is shown able to model extremely heterogeneous clutter, such as that of urban areas, that cannot be properly modeled with K distributions. One of the parameters of this special case is related to the degree of homogeneity, and a limiting case is that of a scaled speckle. The advantage of the G/sup 0/ distribution becomes evident through the analysis of a variety of areas (urban, primary forest and deforested) from two sensors.

Classifying Multifrequency Fully Polarimetric Imagery With Multiple Sources of Statistical Evidence and Contextual Information

This paper presents the use of a new distribution for fully polarimetric image classification. Several classification strategies are compared in order to assess the importance of a careful statistical modeling of the data and the complementary nature of the information provided by different frequencies. Spatial context, which is relevant in order to obtain good results with noisy data, is described by means of the multiclass Potts model, and an iterated conditional modes classification algorithm that employs pseudolikelihood is proposed. The data are described using multivariate Gaussian laws and fully multilook polarimetric distributions arising from the multiplicative model. L-band, C-band, and both bands are used to assess the influence of dimensionality on the classification. Contextual and pointwise maximum-likelihood classifications are compared using real data. Results show that both context and number of frequencies contribute for better classification products, and that, a careful statistical description of the data leads to improved results.

Exploratory study of the relationship between tropical forest regeneration stages and SIR-C L and C data☆

Abstract In this article, the relationship between secondary forest regrowth stage and SIR-C SAR data is assessed, for an area located near to the Tapajos National Forest, south of Para State, in the Amazon region. These regeneration stages are mapped by making use of a consecutive annual sequence of Landsat TM (Thematic Mapper) images. Using this map as a mask over the radar images, the tonal means (expressed in dB) and coefficient of variation (CV) for several second-growth succession stages classes are calculated. It is shown that the discrimination between regeneration stages is difficult when individual ″small areas are used, but this discrimination might be possible when L-band means over a “large area are computed. In particular, the LHV band seems to carry more information. The maximum difference of means among classes occurred in this band and it is of about 5 dB. The CV appeared to be less well suited than the mean value for regeneration stage discrimination, although some discrimination among early stages of regeneration may be possible using this measure at L-band.

Hypothesis Testing in Speckled Data With Stochastic Distances

Images obtained with coherent illumination, as is the case of sonar, ultrasound-B, laser, and synthetic aperture radar, are affected by speckle noise which reduces the ability to extract information from the data. Specialized techniques are required to deal with such imagery, which has been modeled by the G 0 distribution and, under which, regions with different degrees of roughness and mean brightness can be characterized by two parameters; a third parameter, which is the number of looks, is related to the overall signal-to-noise ratio. Assessing distances between samples is an important step in image analysis; they provide grounds of the separability and, therefore, of the performance of classification procedures. This paper derives and compares eight stochastic distances and assesses the performance of hypothesis tests that employ them and maximum likelihood estimation. We conclude that tests based on the triangular distance have the closest empirical size to the theoretical one, while those based on the arithmetic-geometric distances have the best power. Since the power of tests based on the triangular distance is close to optimum, we conclude that the safest choice is using this distance for hypothesis testing, even when compared with classical distances as Kullback-Leibler and Bhattacharyya.

Speckle reduction in polarimetric SAR imagery with stochastic distances and nonlocal means

This paper presents a technique for reducing speckle in Polarimetric Synthetic Aperture Radar (PolSAR) imagery using Nonlocal Means and a statistical test based on stochastic divergences. The main objective is to select homogeneous pixels in the filtering area through statistical tests between distributions. This proposal uses the complex Wishart model to describe PolSAR data, but the technique can be extended to other models. The weights of the location-variant linear filter are function of the p-values of tests which verify the hypothesis that two samples come from the same distribution and, therefore, can be used to compute a local mean. The test stems from the family of (h-phi) divergences which originated in Information Theory. This novel technique was compared with the Boxcar, Refined Lee and IDAN filters. Image quality assessment methods on simulated and real data are employed to validate the performance of this approach. We show that the proposed filter also enhances the polarimetric entropy and preserves the scattering information of the targets.

Classification of SAR images using a general and tractable multiplicative model

Among the frameworks for Synthetic Aperture Radar (SAR) image modelling and analysis, the multiplicative model is very accurate and successful. It is based on the assumption that the observed random field is the result of the product of two independent and unobserved random fields: X and Y. The random field X models the terrain backscatter and, thus, depends only on the type of area to which each pixel belongs. The random field Y takes into account that SAR images are the result of a coherent imaging system that produces the well-known phenomenon called speckle noise, and that they are generated by performing an average of n statistically independent images (looks) in order to reduce the noise effect. There are various ways of modelling the random field X; recently the Γ−1/2(α, γ) distribution was proposed. This, with the usual Γ1/2(n, n) distribution for the amplitude speckle, resulted in a new distribution for the return: the (α, γ, n) law. The parameters α and γ depend only on the ground truth, and n is the number of looks. The advantage of this distribution over the ones used in the past is that it models very well extremely heterogeneous areas like cities, as well as moderately heterogeneous areas like forests and homogeneous areas like pastures. As the ground data can be characterized by the parameters α and γ, their estimation in each pixel generates parameter maps that can be used as the input for classification methods. In this work, moment estimators are used on simulated and on real SAR images and, then, a supervised classification technique (Gaussian maximum likelihood) is performed and evaluated. Excellent classification results are obtained.

Improved estimation of clutter properties in speckled imagery

This papers aim is to evaluate the effectiveness of bootstrap methods in improving estimation of clutter properties in speckled imagery. Estimation is performed by standard maximum likelihood methods. We show that estimators obtained this way can be quite biased in finite samples, and develop bias correction schemes using bootstrap resampling. In particular, we propose a bootstrapping scheme which is an adaptation of that proposed by Efron (J. Amer. Statist. Assoc. 85 (1990) 79). The proposed bootstrap does not require the quantity of interest to have closed form, as does Efforts original proposal. The adaptation we suggest is particularly important since the maximum likelihood estimator of interest does not have a closed form. We show that this particular bootstrapping scheme outperforms alternative forms of bias reduction mechanisms, thus delivering more accurate inference. We also consider interval estimation using bootstrap methods, and show that a particular parametric bootstrap-based confidence interval is typically more reliable than both the asymptotic confidence interval and other bootstrap-based confidence intervals. An application to real data is presented and discussed.

Texture in airborne SAR imagery of tropical forest and its relationship to forest regeneration stage

Abstract At C-band, SAR imagery often exhibits little variation in mean amplitude between different types of natural land cover. However, there is frequently a large amount of information to be found in the textural properties of such imagery, especially when it is acquired at high spatial resolution. This textural information may be useful in observing processes that affect the homogeneity of land surface vegetation such as the staged succession of regenerating tropical forest following human disturbance which is characterized by the gradual decrease in canopy homogeneity as regrowth species are succeeded by hardwood species. In this study, three techniques of measuring the texture in C-band airborne SAR imagery from a tropical forest region in central Brazil are compared. The dependence of these measures on the stage of forest regeneration is assessed by using a temporal sequence of Landsat TM imagery to independently estimate the age of regrowth. Each texture measure is able to discriminate well betwee...

Classification of Segments in PolSAR Imagery by Minimum Stochastic Distances Between Wishart Distributions

A new classifier for Polarimetric SAR (PolSAR) images is proposed and assessed in this paper. Its input consists of segments, and each one is assigned the class which minimizes a stochastic distance. Assuming the complex Wishart model, several stochastic distances are obtained from the h - φ family of divergences, and they are employed to derive hypothesis test statistics that are also used in the classification process. This article also presents, as a novelty, analytic expressions for the test statistics based on the following stochastic distances between complex Wishart models: Kullback-Leibler, Bhattacharyya, Hellinger, Rényi, and Chi-Square; also, the test statistic based on the Bhattacharyya distance between multivariate Gaussian distributions is presented. The classifier performance is evaluated using simulated and real PolSAR data. The simulated data are based on the complex Wishart model, aiming at the analysis of the proposal with controlled data. The real data refer to a complex L-band image, acquired during the 1994 SIR-C mission. The results of the proposed classifier are compared with those obtained by a Wishart per-pixel/contextual classifier, and we show the better performance of the region-based classification. The influence of the statistical modeling is assessed by comparing the results using the Bhattacharyya distance between multivariate Gaussian distributions for amplitude data. The results with simulated data indicate that the proposed classification method has very good performance when the data follow the Wishart model. The proposed classifier also performs better than the per-pixel/contextual classifier and the Bhattacharyya Gaussian distance using SIR-C PolSAR data.