Self-Supervised Super-Resolution for Multi-Exposure Push-Frame Satellites

TL;DR

This paper introduces a self-supervised super-resolution method for multi-exposure satellite images that effectively handles noise, exposure inaccuracies, and varying sequence lengths, improving upon existing single-exposure techniques.

Contribution

The proposed approach is the first to address super-resolution of multi-exposure satellite sequences with self-supervision, incorporating noise-awareness and robustness to exposure errors.

Findings

Outperforms adapted single-exposure methods on synthetic data

Effective handling of signal-dependent noise in real data

Robust to exposure time inaccuracies

Abstract

Modern Earth observation satellites capture multi-exposure bursts of push-frame images that can be super-resolved via computational means. In this work, we propose a super-resolution method for such multi-exposure sequences, a problem that has received very little attention in the literature. The proposed method can handle the signal-dependent noise in the inputs, process sequences of any length, and be robust to inaccuracies in the exposure times. Furthermore, it can be trained end-to-end with self-supervision, without requiring ground truth high resolution frames, which makes it especially suited to handle real data. Central to our method are three key contributions: i) a base-detail decomposition for handling errors in the exposure times, ii) a noise-level-aware feature encoding for improved fusion of frames with varying signal-to-noise ratio and iii) a permutation invariant fusion strategy by temporal pooling operators. We evaluate the proposed method on synthetic and real data and show that it outperforms by a significant margin existing single-exposure approaches that we adapted to the multi-exposure case.