Newsroom

Best paper: Carla Geara awarded in IEEE Geoscience and Remote Sensing Letters

20 April 2026
Carla-Geara
The paper « Extending InSAR2InSAR to Sentinel-1 Data » [1], published in IEEE Geoscience and Remote Sensing Letters, receives the IEEE Geoscience and Remote Sensing Society 2025 Letters Prize Paper Award.
This research is conducted by Carla Geara, a PhD student under an industrial-academic partnership with the company Kanop, and supervised by Florence Tupin at Télécom Paris in collaboration with ONERA.

The work builds on previous research by the IMAGES team at LTCI, which has been developing deep learning methods to enhance satellite radar images for several years. These images, acquired by emitting electromagnetic waves in the microwave frequency range, are subject to strong fluctuations (a phenomenon known as « speckle noise »).

In this paper, we focus on interferometric data, which allows height information to be derived from two radar images by exploiting phase differences. The proposed approach extends and adapts the SAR2SAR [2] and POLSAR2POLSAR [3] methods, which processed amplitude (wave magnitude) and polarimetric data. It leverages the specific acquisition geometry of Sentinel-1 images in IW (Interferometric Wide) TOPSAR mode, which provides radar images with overlapping areas viewed from different angles. This results in two acquisitions with independent speckle noise, enabling the training of a neural network in a self-supervised manner without ground truth images.

This research was carried out as part of Carla Geara’s PhD thesis, which explores the potential of radar data for forest monitoring applications developed by the start-up Kanop.

And also: Carla Geara’s testimonial when she pursued an engineering double degree Lebanese University/Télécom Paris

The Geoscience and Remote Sensing Society (GRSS) is a technical Society of the Institute of Electrical and Electronics Engineers (IEEE). GRSS fosters engagement of its members for the benefit of society through science, engineering, applications, and education as related to the development of the field of geoscience and remote sensing.

[1] C. Geara, C. Gelas, L. d. Vitry, E. Colin and F. Tupin, « Extending InSAR2InSAR to Sentinel-1 Data » in IEEE Geoscience and Remote Sensing Letters, vol. 22, pp. 1-5, 2025, Art no. 4008105

[2] SAR2SAR: a semi-supervised despeckling algorithm for SAR imagesEmanuele Dalsasso, Loïc Denis, Florence Tupin, IEEE Journal of Selected Topics in Applied Earth Observations and Remote Sensing, In press, pp.1-1. ⟨10.1109/JSTARS.2021.3071864⟩

[3] POLSAR2POLSAR: A SEMI-SUPERVISED DESPECKLING ALGORITHM FOR POLARIMETRIC SAR IMAGES Cristiano Ulondu Mendes, Emanuele Dalsasso, Yi Zhang, Loïc Denis, Florence Tupin ISPRS Journal of Photogrammetry and Remote Sensing, 2025, 220 (0924-2716), pp.783-798

1. The TOPS acquisition mode of Sentinel-1 (a mission by the ESA – European Space Agency) results in an overlap of the acquired images (burst overlap). The interferogram created from the two images (first and second acquisitions) is therefore observed twice in the burst overlap: once in the "top" view and once in the "bottom" view, with independent noise realizations. These can be used to train a neural network.
1. The TOPS acquisition mode of Sentinel-1 (a mission by the ESA – European Space Agency) results in an overlap of the acquired images (burst overlap). The interferogram created from the two images (first and second acquisitions) is therefore observed twice in the burst overlap: once in the « top » view and once in the « bottom » view, with independent noise realizations. These can be used to train a neural network.
2. Example of results: on the left, the phase and coherence of the original interferogram; on the right, the phase and coherence after processing by the network.
2. Example of results: on the left, the phase and coherence of the original interferogram; on the right, the phase and coherence after processing by the network.