The KaRIn interferometric altimeter, which is the main instrument of the SWOT mission has been designed to map the topography of water surfaces. While its operating principle is similar to that of previous SAR interferometers such as SRTM or TanDEM-X, its characteristics have been adapted to its purpose: near-nadir incidence to get a strong signal on water surfaces at the expense of land surfaces, small baseline, and Ka-Band wavelength. Despite this, SWOT images show a good coherence on certain types of land surfaces such as sand dunes and early results of topographic processing of SWOT data on these surfaces are very promising. For instance, this could enable new ways to monitor dune migration through SWOT’s high temporal revisit and vertical accuracy.

Water detection is a key step in the operational processing of KaRIn HR images from the SWOT mission, which provides water surface extent and elevation for continental water surfaces globally and repeatedly. A new water detection method has been developed for this sensor due to its specific characteristics. In this article we present examples of detected water masks and preliminary assessments of their accuracy. We also provide perspectives for further improvements in the water detection algorithm.

Dam reservoirs account for only a small fraction of the global freshwater but they have major importance for human activities, providing water for agriculture, domestic, recreational, and industrial use, especially in a context of water scarcity induced by climate change [1] . They also have major impacts on the hydro- and ecosystems [2] . While they are subject to numerous studies across multiple disciplinary fields [3] , [4] , no exhaustive database of existing dam reservoirs is available, and the available data are limited to the largest reservoirs. Thus, there is a need for a method to automatically extract dam reservoirs from widely available data such as satellite images. This task is different from the broader water surface extraction task [5] , [6] as the goal is also to distinguish dam reservoirs from other waterbodies. Some approaches have been proposed to achieve it, for example by Van Soesbergen et al. [7] . However, this extraction task using only single images can be very difficult for very small waterbodies (below 1 hectare).

This paper presents a semi-guided method to detect lakes in Sentinel-1 SAR data. The proposed approach is an adaptation of the grab-cut framework developed in [1]. Starting from a coarse bounding box around the lake, an accurate segmentation is extracted using a Conditional Random Field formalism and a graph-cut based optimization. Then an extension of this approach to process jointly a stack of multi-temporal data is presented. A temporal regularization term is introduced to control the joint segmentation. The proposed approach is evaluated on Sentinel-1 datasets. Qualitative and quantitative results demonstrate the interest of the proposed framework and its robustness to the initial-ization polygon of the lake.

The increasing availability of synthetic aperture radar (SAR) time series creates many opportunities for remote sensing applications, but it can be challenging in terms of amount of data to process. This letter discusses the interest of the geometric mean to average SAR time series. First, the properties of the geometric mean and the arithmetic mean are compared. Then, a speckle-reduction method specifically designed to improve images obtained with the geometric mean is presented. This method is based on an adaptation of the MuLoG framework to take into account the specific distribution of the geometric mean. Finally, applications of this denoised geometric-mean image are presented.

This paper presents a despeckling method for Sentinel-1 GRD images based on the recently proposed framework "SAR2SAR": a self-supervised training strategy. Training the deep neural network on collections of Sentinel 1 GRD images leads to a despeckling algorithm that is robust to space-variant spatial correlations of speckle. Despeckled images improve the detection of structures like narrow rivers. We apply a detector based on exogenous information and a linear features detector and show that rivers are better segmented when the processing chain is applied to images pre-processed by our despeckling neural network.

Synthetic aperture radar (SAR) and optical satellite image registration is a field that developed in the last decades and gave rise to a great number of approaches. The registration process is composed of several steps: feature definition, feature comparison and optimization of a geometric transformation between the images. Feature definition can be done using simple traditional filtering or more complex deep learning (DL) methods. In this paper, two traditional approaches and a DL approach are compared. One can then wonder if the complexity of DL is worth to address the registration task. The aim of this paper is to quantitatively compare approaches rooted in distinct methodological areas on two common datasets with different resolutions. The comparison suggests that, although more complex, the DL approach is more precise than traditional methods.