Philippe Courmontagne

Detection of underwater mines is a present crucial strategic task. Most of the methods used to detect and classify objects lying on the seafloor are based on the interpretation of the shadows of the objects, others are focusing on the echo itself. In this paper, we propose to use a one-dimensional detector based on the noisy data projection onto the most convenient basis for an optimal detection. This one is applied on the SAS data line after line, considering as line the data vector in a direction perpendicular to the SAS movement. Results obtained on real SAS data are presented to show the efficiency of such an approach.

SAS (synthetic aperture sonar) has been used in sea bed imagery. Indeed high resolution images provided by SAS are of great interest, especially for the detection, localization and eventually classification of objects lying on sea bed. But as any image obtained with a coherent system, SAS images are highly corrupted by a granular multiplicative noise, called speckle noise which reduces spatial and radiometric resolutions. The appraisal of despeckling methods or in more general denoising methods is a never ending discussion. Since, these techniques are applied on images, one can somewhat merely look at the image and have a good idea of the general outcomes. Yet, in practice this would take too much time and overall would not be objective at all. Thus, people doing research work on filter wanted to assess their developed methods and tried to find out objective assessing processes. In this paper, we present some of the scientific literature proposals and, by the way of experiments on real data, discuss about them.

Detecting useful information in the underwater domain has taken an important place in many research works. But this problem is not easy to solve because of the large diversity of acoustic sources in the underwater medium. To illustrate this point, let us consider separately two main kinds of information. On the one hand, artificial signals coming from human activities (communication, imaging, detection, ...) and, on the other hand, natural signals due to biological or seismic activities. This non exhaustive enumeration shows that underwater information is very diverse and because all entities live together, the information coming from each one is necessarily interacting with all the others. So one can easily imagine that numerous interaction problems between these two kinds of information arise. In this paper, we propose to answer the problem of signal detection in noisy environment using an optimal subspace projection. The proposed method is confronted to real data, with signals resulting from underwater acoustic measurements: several whale echoes (the signal of interest to be detected) and a superposition of various marine acoustic signatures.

SAS (Synthetic Aperture Sonar) has been used in sea bed imagery. Indeed, high resolution images provided by SAS are of great interest, especially for the detection, localization or eventually classification of objects lying on sea bed. But, SAS images are highly corrupted by a granular multiplicative noise, called speckle noise, which reduces spatial and radiometric resolutions. For this reason, an automatic analysis of these images is not so evident. A solution can consist on the use of a filtering before process, without a spatial resolution degradation. The purpose of this article is to present a new process consisting on the jointly use of the stochastic matched filter and an autoadaptive mean filter. Furthermore, in order to well preserve the spatial resolution, we propose to use as a criterion for the stochastic matched filter the minimization between the speckle noise local statistics with the removal signal ones, allowing a subimage size adaptation. Results obtained on real SAS data are proposed and compared with those obtained using another stochastic matched filtering based denoising method.

Detection and classification of underwater mines with synthetic aperture sonar (SAS) images is a challenge that can be performed in studying either the echoes or the shadows of mines. But, as any images obtained with a coherent system, SAS images are highly corrupted by the speckle noise, which reduces spatial and radiometric resolutions. So such a noise can be very disturbing for the interpretation and the automatic analysis of SAS images. To reduce the speckle level, filtering methods are generally used but all of them strongly deteriorate either the shadow or the echo of the mine. The purpose of this article is to compare several stochastic matched filter based denoising methods, in order to determine which of them is the most appropriate to enhance both echoes and shadow mines. Results obtained on real SAS data are presented and discussed.

SAS images are highly corrupted by a granular multiplicative noise, called the speckle which reduces spatial and radiometric resolutions. The purpose of this article is to present a new adaptive processing that allows speckle noise reduction in SAS images, while keeping the spatial resolution. This processing is based on the two-dimensional stochastic matched filtering method with a signal autocorrelation function adaptation and a sub-image size adaptation. Comparison with some classical approaches on real SAS data reveal the efficiency of such an idea

Detecting straight patterns like ship wakes on a SAR image is not easy because there is no a priori information on orientation and position, moreover, SAR images are speckle noised. This article describes a ship wake detection technique based on the discrete Radon transform and stochastic matched filtering (SMF) used in detection. The association of these two processing methods leads to a detection algorithm that only requires the knowledge of the second order statistics of the signal and the noise. Experimentation on real SAR images shows the efficiency of the technique

Over the past few years a great deal of research has been dedicated to SAS image de-noising. Indeed, such images are highly corrupted by a granular multiplicative noise, called the speckle noise, due to the construction of the image itself. This noise reduces spatial and radiometric resolutions. The goal of the research in the area of SAS image de-noising is double: on one hand to reduce the noise level and on the other hand to preserve the spatial resolution. In this context, recently a new approach has been developed: the curvelet transform. Such a method allows good spatial resolution preservation but the results are still slightly disturbed by the introduction of disturbing straight lines due to the process itself. The purpose of this article is to improve the de-noising power of the curvelet transform, while keeping its spatial resolution preservation. To achieve such an objective, the idea is to couple this transform with a filtering technique, known as the stochastic matched filtering method. Results obtained on real SAS data are presented and compared with those obtained using the classical curvelet transform.