The satellite jitter effect observable in the digital elevation model of difference, herein referred to as DoD, generated using tri-stereo pairs of Pléiades images is a common phenomenon that can reduce the precision of elevation measurements making it challenging for 3D change detection in the land surface. To address this issue, previous studies used correction methods based on polynomial and sinusoid fitting to reduce the jitter effect. However, such approaches do not entirely remove these undulations. In this paper, the image noise reduction task is framed as a signal filtering problem, to eliminate these repetitive patterns while preserving relevant information for accurate change detection using Fourier transform. The noise frequency is identified through the frequency spectrum using a threshold-based method, allowing the subsequent application of a finite impulse response filter to remove the noise. Experiments conducted on Pléiades DoDs data covering the Lebna watershed located in the north of Tunisia validate the effectiveness of this approach. The results demonstrate that the Fourier transform-based filtering method significantly outperforms the state-of-the-art methods, both in terms of qualitative visual assessment and quantitative performance metrics.

The Lebna watershed in northeast Tunisia is marked by its important agricultural activities and its hydro-geological settings, enhancing its susceptibility to various geohazards including land subsidence and uplift. This study monitors the Line of Sight (LOS) deformation through the Interferometric Synthetic Aperture Radar (InSAR) technique. The main observations are LOS velocity maps derived from the Small Baseline Subset - InSAR, which were calculated from ESA Sentinel-1 satellites for the period 2015–2023. The SAR data resulted in velocities ranging between $$-1.05$$ and $$+0.80$$ cm/year. Both subsidence and uplifting trends were observed in different areas, indicating that the ground deformations in the study region could be dependent on geological and hydrogeological factors.

In this study, we explore the relationship between ground subsidence acceleration and precipitation fluctuations in the Cap-Bon peninsula in northern Tunisia. Our findings reveal that this phenomenon shows uneven trends in relation to the terrain, as well as the types of vegetation and land use in the area. The analysis of the Interferometric Synthetic Aperture Radar (InSAR) ground subsidence deformation time series is considered within two critical periods for the region between 2018-2019 and 2021-2022. Results reveal distinct non-uniform subsidence in Cap-Bon peninsula, with more pronounced effects observed in the irrigated area of Zaouiet Jedidi, reaching a maximum subsidence rate of -19.62 cm/year. This area is known to be protected because of its importance for the agricultural production of citrus fruits on a national level. Then, after a thorough examination of the factors contributing to subsidence, it has been established that the uneven deformation is mainly due to the over-exploitation of groundwater within a relatively dry period.

This paper investigates the relevance of the time series land subsidence analysis to assess the water level changes, both at the surface and in the aquifer, in the coastal study area in the Cap-Bon peninsula in northern Tunisia. We used the advanced Parallel Small BASeline Subset (P-SBAS) InSAR (Interferometric Synthetic Aperture Radar) processing chain running on the Geohazards Exploitation Platform (GEP) to generate a deformation time series maps from Sentinel-1 (S-1) interferometric wide (IW) swath data sets. These data were acquired in 2017-2022 over the study area, and comprising 154 scenes in descending mode. Results showed that the subsidence in the Cap-Bon peninsula, that was related to groundwater overexploitation, with the highest average land subsidence occurring in 2022 (9 cm) and the lowest in 2017 and 2020 (< 1 cm) is proportional to the water surface extent and then to the water quantity on the surface of the watersheds. The water surface extent was jointly delimited using the InSAR coherence maps and corresponding NDWI (Normalized Difference Water index) computed over Sentinel-2 (S-2) data.

Ensuring the authenticity of scanned documents is of major concern, these ones being often admitted as evidence by organizations. “Is there any way to verify that a document was scanned by a device without having access physically to the source device itself?” is a wide-open question. In this paper, we aim at answering it in the affirmative by means of the first data-driven hybrid machine learning framework that compares image noise features to check if two documents have been digitized with the same scanner or not. Such a problem is known as device linking function. Different comparative experiments conducted on the same and different scanner models on a broad set of administrative documents demonstrate that our method is efficient in linking scanned images even if scanner devices are unknown to the investigator. Our success rate of 96% appears to be the novel state of art reference in such application domain.

The minimum mean square error (MMSE) based filters constitute an important branch of speckle reduction filtering techniques of synthetic aperture radar (SAR) and Polarimetric SAR (PolSAR) images. In this paper, the performance of MMSE based in terms of speckle reduction and spatial detail preservation is revised. To preserve the spatial details, the weight coefficient$b$should be equal to 1. However, the authors demonstrated in this paper that this parameter do not fulfill this condition. As a consequence, the MMSE-based filters have some limitations in spatial detail preservation. In PolSAR filtering, the weight coefficient$b$is estimated from the span image. However, it has been demonstrated that the standard deviation of the span (i. e.$\sigma_{v(span)}$) is a pixel dependent. The use of a fixed$\sigma_{v(span)}$for the entire image as applied in the majority of the state of the art MMSE filters is not optimal.

The use of the span as a reference image in PolSAR speckle filtering is investigated. It has been demonstrated in this study that span reduced the speckle noise in extended homogeneous areas differently, degraded the spatial details and distorted the filtered intensity image. Then, the impact of span on the filtered PolSAR images is evaluated. The results showed that the drawbacks of the span filtering in terms of speckle filtering, spatial detail and polarimetric information preservation are propagated to the minimum mean square error MMSE-based PolSAR filters. The estimation of the filtering weights from the intensity channels rather than the span resolved these issues.

The ease of use and the capabilities of image editing tools has raised the challenges in the emerging field of digital image forensics related to scanned documents. Unfortunately, the universality of current methods and their applicability in real world scenarios have not been proven yet due to the absence of a standardized image database. In this paper, we introduce a novel test collection of more than 4500 images annotated with respect to each scanner as a useful tool for forensics in-vestigators to test and compare scanner-based forensic techniques. It is an image database that contains document of various content scanned with more than one resolution with 11 different scanner instances of widely known brands. This selection is based on our latest work on the identification of scanners at the origin of digitized documents and is adapted to fit any source scanner identification technique. The SUPATLANTIQUE database is available for free to the research community and is intended to become a useful and a reference resource for researchers in this field.

Multi-temporal techniques of differential interferometry are providing accurate performance in many thematic applications of soil displacements monitoring. In this context, the analysis of the soil movements that happen after flooding events is crucial for the natural hazard management task. In this paper, we are considering the parallel Small BAseline Subset (P-SBAS) technique using Sentinel-1 dataset in order to produce a multi-temporal study of the soil displacement that happened in 2018 after flooding in Tunisia. Moreover, we are performing a qualitative comparison between the generated mean velocity map of the displacement and the lithofacies map that describes the geomorphological characteristics of the soil in the study area. The achieved results of two test sites in the north of Tunisia show an effective connection between the mapped areas of displacement using P-SBAS and the regions that are declared by the ground-truth map as prone to post-flood movement.

In this paper, the application of the synthetic aperture radar (SAR) infinite number of looks prediction (INLP) filter is extended to polarimetric SAR (PoISAR) speckle filtering. The scalar linear regression rule has been adapted to PolSAR context in order to preserve the polarimetric information. Experimental results using simulated and airborne PolSAR data show that the proposed approach improved the polarimetric filtering criteria.

In this paper, we enhanced the performance of the iterative minimum mean square error filter (IMMSE) in terms of speckle suppression and detail retaining by establishing another version of the minimum mean square error MMSE estimator and adapting the iterative procedure to the non-local structure similarity. The non-local principle is introduced in the calculation of the IMMSE filter parameters. Simulated and Spaceborne (Sentinel1 Amdoun region-Tunisia) high-resolution synthetic aperture radar SAR data were used for validation.

In this paper, we extend the use of the synthetic aperture radar (SAR) iterative minimum mean square error (IMMSE) technique to polarimetric SAR (PolSAR) speckle filtering. Experimental results prove that the proposed approach especially reduces speckle in the extended homogenous areas and, meanwhile, better preserves the spatial details and the polarimetric information, compared with other classical methods.

The aim of this paper is to study the potential of synthetic aperture radar (SAR) for estimating wheat height using interferometry and polarimetry. Our contribution in this paper consists on a statistical relationship that we establish between field height measurement and backscattering coefficient, polarimetric parameters and interferometric coherence. For experimental validation, the height of wheat crops was determined through the different phenological stage for two types of wheat conducted irrigated and rainfed wheat. In the same period of field campaign, eight Radarsat-2 images and seven Sentinel-1 images were acquired over the region of Sidi Bouzid in the center of Tunisia. From these images, temporal variation of backscattering coefficient, interferometric coherence and polarimetric parameters were extracted and compared to the height variation of wheat. The results showed a weak correlation between wheat height and backscattering coefficient. A high correlation between height of wheat and polarimetric parameters principally Radar vegetation index (RVI) and Pedestal height ( r=0.9 and r=0.92 respectively) and an acceptable correlation between height and interferometric coherence (r=0.6).

In this paper, we are mainly interested in identifying and measuring the ground subsidence that occurred after the flooding of Oued EI Maadin tributary in Tunisia. For this purpose, we are adopting the Differential Interferometry Synthetic Aperture RADAR approach (D-InSAR) using a set of multi-temporal Sentinel1 SLC data. Furthermore, we are considering an offset tracking technique that provides assessments of the land subsidence movements. To do so, we are using two Sentinel1 GRD images that were acquired before and after the soil deformation event.

The filtered pixels of the infinite number of looks prediction (INLP) filter are determined by the application of a linear regression between means and variances for various window sizes. In this paper, we improved the performance of the INLP filter in terms of versatility, computing complexity, speckle reduction and spatial detail preservation by adapting the parameter Nmin which is the key parameter to scene variability. Results show that the improved INLP outperformed the improved sigma filter, the non-local filter and the original INLP filter.

In this paper we propose a semi-automatic off-roads and trails extraction method using synthetic aperture radar (SAR) data from Sentinel-1 satellite. In fact, these categories of rural and desert roads, very solicited on the Tunisian borders by smugglers, are often changing and need more close monitoring. Thus, we propose here to take benefit from free SAR Sentinel-1 data with high resolution and repetitive coverage. Taking into account the particularities of these roads, having curvilinear features, and those of the data, degraded with granular noise that inherently exists in SAR, we propose a more general approach based on anisotropic diffusion and dynamic snakes adapted to the case of off-roads and trails. Application of the developed approach to the Sentinel 1 SAR data acquired over the Tunisian frontier lead to the extraction of the appropriate road edges.

In this paper, we exploited the minimum mean square error (MMSE) estimator to define an iterative synthetic aperture radar (SAR) speckle filtering process. Hence, by optimally choosing the window size and the number of iterations, the proposed iterative filter outperforms classical MMSE speckle filtering techniques such as the improved Lee filter in terms of speckle reduction and spatial detail preservation.

Speckle filtering in synthetic aperture radar (SAR) images is essential for the extraction of significant information for homogeneous extended targets. In this letter, we exploited the minimum mean square error to establish a linear rule between the values of the filtered pixels and their variances. Then, the filtered pixel for infinite number of looks (INL) was predicted by linear regression of means and variances for various window sizes. The results show that the proposed INL-prediction filter improved the filtering performances of the original filters. Simulated and real SAR data were used for validation.

Leaf area index (LAI) is very used to reveal the vegetation situation. To estimate the LAI for cereal, both direct and indirect methods have been used. In particular, remote sensing is a fast and reliable technique to develop the LAI estimation models. In this work, we present the potential of Sentinel 1 images for cereal estimation LAI in the center of Tunisia under semi arid climate. We established a statistical relationship between field LAI measurement for irrigated and rainfed wheat and backscatter coefficient based on an empirical analysis. This will be very useful in order to predict the water stress in a subsequent step. For experimental validation, the LAI of wheat crops were determined through the crop growth stages using eight Sentinel-1 images. The results showed a significant correlations, for the irrigated wheat, between LAI and backscatter coefficient for VV polarization (r values of -0.7) and for HV (r value of -0.5). For the rainfed wheat only the VV polarization showed a significant correlation (r=0.6).

This work deals with unsupervised classification approach for mapping the flood impacts using interferometric synthetic aperture radar (InSAR) images. In order to estimate the flooding extent, we went towards considering the interferometric coherence information in view of the fact that flooded regions present a loss of coherence between two image acquisitions before and after the flooding. In this paper, we are adopting a fuzzy clustering algorithm to discriminate the flooded and dry areas in our region of interest. The definition of the input attribute vector was developed in [1], where features from InSAR amplitude and coherence images were combined into a single 3D space in which every pixel is represented by three-values vector (master image, slave image, coherence image). Then, an improved multichannel fuzzy C-means (FCM) clustering algorithm is developed in order to perform the classification process. Our main contribution is the optimization of the similarity measure of the FCM clustering algorithm using the InSAR coherence map. Experimental results and quantitative evaluation are given considering the Envisat data acquired over the Mellegue river in the north of Tunisia in 2005.

This paper deals with the interferometric synthetic aperture radar (InSAR) phase unwrapping problem. The proposed appraoch is based on the relationship between the gradient vectors of the observed wrapped phase and the true phase respectively, when the Itoh condition is satisfied. Since this relationship is violated by the residu pixels in the observed wrapped phase, we propose, in this paper, a general problem formulation which takes into account the estimation error due to these residu values. The main idea is based on compressed sensing technique to estimate the true phase and the estimation error simultaneously. To obtain a sparsity representation of the unknown true phase image, we use the curvelet transform orthobasis dictionary and the error is assumed to be a sparse signal. Finally, the minimization of the result l1problem is insured by the Basis Pursuit algorithm (BP). The proposed unwrapping approach shows high performance when tested and validated with simulated and real interferograms and compared with two recent similar algorithms.

This paper presents a new integrated InSAR phase filtering and unwrapping method based on a Markov Random Field model. This approach aims to estimate the noise free unwrapped phase from the observed noisy interferogram. The phase image is modeled using MRF where the correponding energy function is defined. This fuctional contains two parts: the first part is dedicated to the interferogram filtering process and the goal of the second part is to unwrap this filtered phase. The filtered unwrapped phase image is estimated by minimising the proposed energy function. In this paper, we used the genetic algorithm to attempt this minimum. The proposed approach is tested and validated on simulated and real interferograms generated from Envisat satellite acquired over the region of Mahdia in Tunisia.

In This paper, a more simple approach for radar shadow extracting from SAR images is proposed. It's based on 1D geometry projection using a digital elevation model (DEM) and considering the corresponding georeferenced SAR image. The DEM to be processed is firstly rotated into the radar geometry so that each row would be suitable for a radar line of sight. Then, the extraction of shadow would be done per row. The proposed method is tested on DEMs as well as on a simulated hilly modeled by Gaussian functions. The obtained results are very interesting, especially for simulated data with high resolution.

Road network extraction from digitized map consists in partitioning a map into two different classes (road and background) distinguished with their intensity values and geometric shape. In this paper, we propose a semi-automatic method where user intervention is minimal. We adopt a variational framework where a minimization energy is proposed. Roads are tracked by active contours which evolve according to evolution equations derived from the energy minimization. We consider that the energy is composed of three main terms which have never been combined and used in such application. The first term is the geodesic active contour (GAC) which is contour-based and which evolves toward high gradients in the image. The second term matches the intensity distribution inside the contour with a model distribution. The third term is a geometric constraint which describes the thin structure of routes. The user intervention is restricted to learning the road intensity model.

Soil salinization is one of the most hazardous phenomenon accelerating the land degradation processes. Map ping and tracking soil salinity changes is fundamental for anticipating natural disaster, such as desertification, in arid and semi-arid regions. In this work, we establish an empirical model for soil salinity mapping based on a gaussian mixture and using field electrical conductivity (EC) measures. The developed model is tested on saline soil samples collected from the semi-arid region of Kairouan located in central Tunisia. It is based on statistical moments derived from multiband (HH and VV) intensity synthetic aperture radar (SAR) data of the Envisat satellite. The resulting salinity map is composed of three classes of salinity (Low, Medium and High) with respect to the EC measurements. The developed model is validated for low salinity distribution, whereas, it needs more samples to be generalized for medium and high soil salinity content.

Interferometric Synthetic Aperture Radar (InSAR) measurements are often biased due to atmospheric effects. Especially, the tropospheric water vapor engenders a delay of SAR signal propagation. In this paper, we propose a specific methodology for atmospheric effects correction on SAR interferograms. It is based on ancillary data collected from NOAA-AVHRR sensor. The specificity of the approach consists in its applicability where no ground truth GPS measurements are available neither for calibration nor for result validation. An adaptive validation demarche is also proposed.

In the literature, several works are led around the radar images especially the detection of the cartographic objects, the 3D reconstruction and the change detection. Concerning this last application, several techniques compete to ensure the best possible result. In this paper, we aim first at developing an automatic detection procedure to compare between similarity measures. Then we propose a change detection technique based on the fusion of two similarity measures. The first one is the Contrast (C) measure [1] and the second one is the Rayleigh Distribution Ratio (RDR) measure [2]. The proposed method has been validated on simulated data and then applied on three radar images.

In this work we demonstrate the potentiel of the InSAR coherence in classifying regions threatened by the desertification. We propose to combine features from InSAR amplitude and coherence images into a single 3D space in which every pixel is represented by its three-values vector (master image, slave image, coherence image) then to apply the multichannel Fuzzy C-Means (FCM) clustering algorithm. Our motivation behind this proposition lies first in the fact that correlation between both of InSAR images can reduce considerably effects of the noise while preserving image details. The second fact is that coherence represents a complementary information to the SAR image amplitude that can well indicates the spatial homogeneity behavior of each contextual class. Two ERS-1 SAR images over the region of Ben Guerdane, in the South-East of Tunisia, are considered for validation.

Fully polarimetric SAR (POL-SAR) images provide a large amount of information through the four channels HH, VV, HV and VH. They proved to be useful in many applications such as delimiting homogenous areas. Such large amounts of data require robust processing algorithms with minimal supervision and low complexity, especially for classification purposes. Most existing classification algorithms (H/A/alpha for instance) use combinations of some or all the channels as features for classification. In this paper, we present a new approach for polarimetric images classification. We are interested in the information of the inter-channel polarime-tric coherence as a feature element for the classification algorithm. The coherence information is known for being used in multi-temporal acquisitions for its advantage of detecting changes in the scenes during time. It is commonly used in in-terferometry. We want to profit from this information in the case of polarimetric images in order to take advantage of the multi-channel property of the data. This coherence classification approach (reading images, computing the coherence and the classification) will be implemented within the OTB (ORFEO ToolBox), the free software which is dedicated especially for remote sensing imagery processing. The approach is tested using images acquired by the CV-580 airborne near Ottawa, Ontario, Canada.

In this paper, we propose a modified filtering algorithm to the Lopez and Fabregas [6] noise reduction algorithm for the interferometric phase noise in SAR interferometry using a multiresolution approach. Our contribution to the existing algorithm consists on the exploitation of the InSAR coherence map in order to generate a more adaptive mask for each decomposition level. The developed algorithm is then tested and validated on simulated and real (ERS, SLC) data.

In this paper, we propose a new Bayesian estimation-based algorithm for two dimensional phase unwrapping of discontinuous phase fields with noisy principal values. The proposed algorithm uses the Markov random field models to build the prior distribution, so the unwrapping problem became equivalent to a minimization of an energy function. Our main contribution in this work is to propose a modification to the classical quadratic potential function, which enforces a global smoothness condition, so that the phase jumps, which result from the phase discontinuities, contribution to the energy are mitigated. This was possible throw weighting the classical quadratic potential function by the probability of occurrence of these jumps which decrease when they increase. Theoretically, this probability follows an hypergeometric distribution which can be approximated as a Gaussian one in order to make easier mathematical manipulations. An analytical expression for the minimization automate was derived.

In this work, we classify the interferometric SAR (InSAR) coherence map into three classes using the Bayes’ theorem. The segmentation procedure is performed using a mixture modelling of the coherence map. The multimodal density of the mixture comprises three component functions characterizing different land surface categories (lake, bare soil, urban …). This work is an ameliorated segmentation approach of that published by the authors in [1]. We test the performance of the proposed mixture model on a dataset about regions with different geophysical characteristics and different time interval between the acquisitions. The results of this study could be used as a supervised learning step for an automatic land cover classification algorithm. This new method classifying the image considering the corresponding InSAR coherence map is particularly powerful for the detection of layover and shadow regions.

Coherence magnitude is a fundamental parameter for the analysis of applications using interferometric synthetic aperture radar (InSAR). The coherence magnitude estimators are biased and need bias removal. The sample coherence magnitude estimation, computed on a window basis, depends on the number of independent samples and theoretical coherence. It has been shown that the sample coherence magnitude estimator is the maximum-likelihood one. It is a biased estimator, especially for low coherence values. In this paper, we present a novel coherence magnitude estimator obtained from the method of moments using "second kind statistics". Classical methods (with regular statistics) for coherence computation are based on a probability density function (pdf) model for estimating regular moments (first kind statistics) defined with the Fourier transform. The proposed approach is based on the same pdf model to compute the second kind statistics defined with the Mellin transform (log-moment). Thus, it is shown that the estimated coherence given by the first log-moment is less biased. Moreover, it is shown that the coherence magnitude estimation from complex coherence maps (interferometric data) using second kind statistics is the optimal estimation procedure of interferometric coherence. It gives the smallest bias near zero comparing with existing estimators. The developed estimation approaches have been applied to obtain coherence measurements from tandem European Remote Sensing 1 and 2 satellite interferometric data, collected over varying terrain with a variety of ground cover types (agriculture field, forest, lake, urban area, sea) in Tunisia, France, and Nepal

The sample coherence magnitude estimation, computed on a window basis, depends on the number of independent samples and theoretical coherence. Classical methods for sample coherence computation are based on probability density function (pdf) model for estimating regular moments (first kind statistics) defined with the Fourier transform. The proposed approach is based on the same pdf model but for computing "second kind statistics" defined with the Mellin-transform. The result performances of this new log-moment (based on the Mellin-Transform) estimator was presented by the authors in R. Abdelfattah and K.M. Nicolas (2003). This paper presents a generalization of the second kind statistics coherence magnitude estimation approach from a complex coherence maps such as interferometric synthetic aperture radar (InSAR) data. The new developed algorithm is much more less biased than existing ones.

Interferometric coherence is a fundamental param- eter which decides on the exploitability of the SAR (Synthetic Aperture Radar) interferogram. The purpose of this paper is to analyse the unbiased complex coherence estimation (1) considering measurements derived from different ERS1/ERS2 interferometric data. The data set was collected over varying terrain with a variety of ground cover types (agriculture field, forest, lake, urban area, sea) in Tunisia, France and Nepal. For a quantitative analysis of the unbiased complex coherence estimation, we compare the obtained results with those given by the sample coherence magnitude estimator. A sensibility analysis with respect to the landscape and the ground cover characteristics and number of independent samples approve the performance of the complex estimator.

We present an optimization of the estimated interferometric synthetic aperture radar (InSAR) coherence, that we had presented in IGARSS'01, which is based on a binary mixture modelisation of the coherence images. The proposed approach is based on a second kind statistics defined as the Mellin-transform of its probability density function (pdf). The Mellin transform of coherence pdf is then computed analytically. It will serve for interferogram parameters estimation by zone growing. Results are validated on a test zone of highly energetic relief (Mustang).

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.

In repeat-pass synthetic aperture radar (SAR) interferometry, the approximations, allowing the phase-to-height conversion, prevent high-resolution mapped relief. In this paper, we present a more general and exact formulation giving a new relationship between the interferogram phase and the target height. It is based on the interferometric SAR geometry and on a better expansion of the path length difference between the sensor and the target. This formulation emphasizes the impact of baseline uncertainties on digital elevation model (DEM) accuracy. A quantitative assessment of the required baseline accuracy is computed. The impact of orbital parameters, used in the new formulation, on interferogram generation is studied. Thus, a new simulator algorithm is developed and tested on a set of reference and simulated DEM examples. Examples of interferogram simulation and DEM generation validate this new approach in the case of a mountainous area in Mustang (Nepal).