Showing papers by "Florence Tupin published in 2002"
TL;DR: The proposed method, which is an adaptation of previous work to the specific case of urban areas, uses the clique potentials of the Markov random field that extracts the road network and a multiscale framework is used.
Abstract: This paper deals with the automatic extraction of the road network in dense urban areas using a few-meters-resolution synthetic aperture radar (SAR) images. The first part presents the proposed method, which is an adaptation of previous work to the specific case of urban areas. The major modifications are 1) the clique potentials of the Markov random field that extracts the road network are adapted and 2) a multiscale framework is used. Results on shuttle mission and aerial SAR images with different resolutions are presented. The second part is dedicated to road extraction combining two SAR images taken with different flight directions (orthogonal and antiparallel passes), and the obtained improvement is analyzed.
119 citations
24 Jun 2002
TL;DR: This article proposes to use a new approach based on "second kind statistics" for solving the problem of binary additive mixture of Gamma law, J. M. Nicolas (2001), and to apply the results to SAR image processing.
Abstract: SAR images are classically analyzed with the help of Goodman approach and multiplicative noise. By this way, speckle is modeled by a Gamma law (for intensity images). A new approach based on "second kind statistics", J. M. Nicolas et al., (2000), identifies multiplicative noise as a "Mellin convolution", J. M. Nicolas et al., (1998), yielding oversimple expression when texture is not homogeneous. In this article, we propose to use this new approach for solving the problem of binary additive mixture of Gamma law, J. M. Nicolas (2001), and to apply the results to SAR image processing.
48 citations
24 Jun 2002
TL;DR: Two other criteria derived from the mean square error analysis are proposed and studied first through their distributions computed using simulated data, and secondly when applied on synthetic and real SAR images.
Abstract: The aim of this paper is to study the use of cross-correlation for radargrammetric applications. Two other criteria derived from the mean square error analysis are proposed. These three criteria are studied first through their distributions computed using simulated data, and secondly when applied on synthetic and real SAR images. Besides, the influence of the use of logarithm or averaged data is studied.
17 citations
24 Jun 2002
TL;DR: This paper discusses an extension of the Symmetric Phase Only Matching Filtering to cover the FMI descriptors of two interferometric SAR images, and tests the method on two pairs of InSAR data in France and Tunisa.
Abstract: The problem of interferometric SAR image coregistartion is addressed. For classical images, the application of the Symmetric Phase Only Matching Filtering (SPOMF) to the Fourier-Mellin Invariant (FMI) descriptors allows an accurate and efficient registration of translated, rotated and scaled images. This paper discusses an extension of the technique to cover the FMI descriptors of two interferometric SAR images. This method is tested on two pairs of InSAR data in France and Tunisa. The results are compared with those of classical cross-correlation registration techniques.
15 citations
01 Jan 2002
TL;DR: In this article, adapted wrapped phase filtering greatly improves the retrieval of tropospheric effects, and the filtered interferograms are then used to model these artefacts and compared to quantify the improvement.
Abstract: Tropospheric inhomogeneities can form a major error source in DinSAR (Differential SAR Interferometry) measurements used in slow deformation monitoring. Previous studies introduced techniques to correct these artefacts. In [1] they propose to evaluate and correct tropospheric effects directly from raw differential interferograms by estimating the phase/altitude correlation. Since the wrapped phase noise in these interferograms influences the correction of tropospheric artefacts its removal is mandatory. In this paper, we aim to show that adapted wrapped phase filtering greatly improves the retrieval of tropospheric effects. The filtered interferograms are then used to model these artefacts. Filtered and unfiltered results are compared to quantify the improvement.
6 citations
10 Dec 2002
TL;DR: A method for learning local potentials using "congregation" of neural networks and supervised learning for handling complex labelling problems driven by local constraints is developed.
Abstract: This paper introduces an approach for handling complex labelling problems driven by local constraints. The purpose is illustrated by two applications: detection of the road network on radar satellite images, and recognition of the cortical sulci on MRI images. Features must be initially extracted from the data to build a "feature graph" with structural relations. The goal is to endow each feature with a label representing either a specific object (recognition), or a class of objects (detection). Some contextual constraints have to be respected during this labelling. They are modelled by Markovian potentials assigned to the labellings of "feature clusters". The solution of the labelling problem is the minimum of the energy defined by the sum of the local potentials. This paper develops a method for learning these local potentials using "congregation" of neural networks and supervised learning.