In this paper, we propose a method for moving-source localization based on beamforming output and on sparse representation of the source positions. The goal of this method is to achieve spatial deconvolution of the beamforming, to provide accurate source localization for pass-by experiments. To perform this deconvolution, we use a smooth approximation of ℓ1/ℓ2[1], which is well suited for the recovery of sparse signals. We validate this method on simulated data, and compare it to the DAMAS-MS method [2], one of the classical methods used in beamforming deconvolution.

Sparse representations have been extended to deal with color images composed of three channels. A review of dictionary-learning-based sparse representations for color images is made here, detailing the differences between the models, and comparing their results on the real and simulated data. These models are considered in a unifying framework that is based on the degrees of freedom of the linear filtering/transformation of the color channels. Moreover, this allows it to be shown that the scalar quaternionic linear model is equivalent to constrained matrix-based color filtering, which highlights the filtering implicitly applied through this model. Based on this reformulation, the new color filtering model is introduced, using unconstrained filters. In this model, spatial morphologies of color images are encoded by atoms, and colors are encoded by color filters. Color variability is no longer captured in increasing the dictionary size, but with color filters, this gives an efficient color representation.

In this letter, a review of the quaternionic least mean squares (QLMS) algorithm is proposed. Three versions coming from three derivation ways exist: the original QLMS [1] based on componentwise gradients, HR-QLMS [2] based on a quaternion gradient operator and iQLMS [3] based on an involutions-gradient. Noting and investigating the differences between the three QLMS formulations, we show that the original QLMS suffers from a mistake in the derivation calculus. Thus, we propose to derive rigorously the criterion following the first way, giving the correct version of QLMS. A comparison with the other QLMS versions validates these results on simulated data.

One of the most challenging applications of time-frequency representations deals with the analysis of the signal issued from natural environment. Recently, the interest for passive underwater context increased, basically due to the rich information carried out by the natural signals. Taken into account the non-linear multi-component time-frequency behavior of such signals, their analysis is a complex problem. In this paper we introduce a new time-frequency analysis concept that aims to extract the non-linear time-frequency components. The main feature of this technique is the joint use of time-amplitude, time-frequency and time-phase information. This is materialized by a short-time polynomial phase modeling and the fusion of local information according to the best locally matches of local cubic frequency modulations. Tests provided on real data illustrate the benefits of the proposed approach.

The estimation of the impulse response (IR) of a propagation channel is necessary for a large number of acoustic applications: underwater communication, detection and localization, etc. Basically, it informs us about the distortions of a transmitted signal in one channel. This operation is usually subject to additional distortions due to the motion of the transmitter-channel-receiver configuration. This paper points on the effects of the motion while estimating the IR by matching filtering between the transmitted and the received signals and introduces a new motion compensation method. Knowing the transmitted signal, the ldquoapparentrdquo speed of each propagation path can be estimated using wideband ambiguity function. Indeed, some interference appears in the wideband ambiguity plane because of the multipath propagation. A warping-based lag-Doppler filtering method is proposed allowing us to accurately estimate the IR of the channel.

Two-pass differential synthetic aperture radar interferometry processing have been successfully used by the scientific community to derive velocity fields. Nevertheless, a precise digital elevation model (DEM) is necessary to remove the topographic component from the interferograms. This letter presents a novel method to detect and retrieve DEM errors by analyzing time series of differential interferograms. The principle of the method is based on the comparison of fringe patterns with the perpendicular baseline. First, a mathematical description of the algorithm is exposed. Then, the algorithm is applied on a series of four one-day European Remote Sensing 1 and 2 satellite (ERS-1/2) interferograms.

Empirical mode decomposition (EMD) is a powerful "time-frequency" tool that is used here in the spatial domain to filter out at each instant, short scale wall pressure fluctuations measured by a linear microphone array beneath a boundary layer. A frequency over streamwise wavenumber representation is used to separate acoustic and turbulent energy: it is obtained by a classical spatial-correlogram which is performed either on original signals or on spatial EMD-filtered signals. It is shown how spatial EMD filtering reduces the spread of the convective energy due to the truncation effects and is tuned to improve the separation of the acoustical energy out of the turbulent energy.

This paper describes source depth estimation in shallow water environments using a Horizontal Line Array of hydrophones on the sea bottom. A method based on modal propagation, using mode amplitude modulus, has been proposed by Nicolas [1]. As knowledge of the sign of mode amplitudes can improve source depth estimation, we present a estimator of these signs and integrate them in the source depth estimation method. Results on real data are shown.

This paper presents a version of MUSIC algorithm for linear vector-sensor arrays based on a complexified quaternionic (biquaternionic) modelization of the output three-components vector-signals. A way of computing the eigenvalue decomposition of a biquaternion valued matrix is introduced and the subspace decomposition of the biquaternionic spectral matrix of the observations is used to define the biquaternionic MUSIC estimator (BQ-MUSIC). Performances of the BQ-MUSIC are compared with classical long-vector technique

This study proposes a matched time-frequency representation construction methodology based on physical model of propagation on underwater environment. The main objective is to filter propagation modes in the time-frequency plane. Optimal representations are then built and test on real dataset for two classical waveguide model

As multicomponent processing is a challenge in many fields such as geophysics, seismology, remote sensing, or electromagnetics, a method that could be used to filter multispectral, multicomponent, or multidimensional data is proposed. In this paper, the method, called multicomponent wideband spectral-matrix filtering (MC-WBSMF), is applied on geophysical data to separate interfering wavefields. The technique is based on the decomposition of a special multicomponent spectral matrix and could extract a given wavefield from a multicomponent dataset. MC-WBSMF is compared to two classical single-component seismic methods (the 2D-Fourier transform or f-k filter and the Radon transform) and to a fully multicomponent method based on singular value decomposition. Finally, MC-WBSMF is tested on a real geophysical subsurface dataset in order to remove surface waves and to enhance reflected waves

The paper deals with seismic signal processing. We propose a new algorithm which performs the dispersion filter estimation using a linear array of geophones. This estimation enables a robust characterization and extraction of the dispersive propagating waves from a seismic profile. The novelty of our method is the use of a simultaneous time delay and phase shift correction in order to estimate the dispersion filter. The resulting algorithm is semi-automatic and requires very few input parameters. The main advantage of our algorithm is that the signals are analysed in the time-frequency domain using all sensors simultaneously while classical approaches are either multi-sensor in the frequency domain or single sensor in the time-frequency domain. Validations on synthetic and real data show the reliability of the algorithm.

The problem of separation of instantaneous mixtures of narrowband signals impinging on sensors often arises in signal processing. In geophysics, the aims of signal processing are the separation and the identification of waves or sources to get a better understanding of the onshore. For a couple of years, multicomponent sensors are used in acquisition and allow one access to a physical property: the polarization. This paper shows that, taking the advantages from the multicomponent sensors, it is possible to have complete wavefield separation without a priori information. We propose to perform the estimation of wave's polarization using an algorithm derived from the Jointly Approximated Diagonalization of Eigen-elements (JADE) technique, which is based on the properties of the 4th order cumulants. From a real seismic acquisition, we present results where Rayleigh wave separation is performed. We also present the analysis of physical properties of the surface waves.