Self-supervised learning of deep despeckling networks with MERLIN: ensuring the statistical independence of the real and imaginary parts - Equipe Image, Modélisation, Analyse, GEométrie, Synthèse Access content directly
Preprints, Working Papers, ... Year : 2023

Self-supervised learning of deep despeckling networks with MERLIN: ensuring the statistical independence of the real and imaginary parts

Abstract

Due to the wide variety of sensors, with different spatial resolutions, operating frequency bands, as well as acquisition modes (Stripmap, Spotlight, TOPS...), despeckling neural networks trained on a given type of SAR images do not perform well on other kinds of images. By considerably simplifying the building of training sets and directly including images from the sensor and acquisition mode of interest, self-supervised learning is a very appealing solution. This paper analyses the preprocessing requirements of the MERLIN strategy that assumes statistical independence of the real and imaginary parts of single look complex SAR images to perform the self-supervised training.
Fichier principal
Vignette du fichier
EUSAR_2024_MERLIN.pdf (3.81 Mo) Télécharger le fichier
Origin : Files produced by the author(s)

Dates and versions

hal-04245667 , version 1 (17-10-2023)

Identifiers

  • HAL Id : hal-04245667 , version 1

Cite

Emanuele Dalsasso, Frédéric Brigui, Loïc Denis, Rémy Abergel, Florence Tupin. Self-supervised learning of deep despeckling networks with MERLIN: ensuring the statistical independence of the real and imaginary parts. 2023. ⟨hal-04245667⟩
53 View
38 Download

Share

Gmail Facebook X LinkedIn More