Self-Supervised Multi-Image Super-Resolution for Push-Frame Satellite Images

Ngoc Long Nguyen, Jeremy Anger, Axel Davy, Pablo Arias, Gabriele Facciolo; Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2021, pp. 1121-1131

Abstract


Recent constellations of optical satellites are adopting multi-image super-resolution (MISR) from bursts of push-frame images as a way to increase the resolution and reduce the noise of their products while maintaining a lower cost of operation. Most MISR techniques are currently based on the aggregation of samples from registered low resolution images. A promising research trend aimed at incorporating natural image priors in MISR consists in using data-driven neural networks. However, due to the unavailability of ground truth high resolution data, these networks cannot be trained on real satellite images. In this paper, we present a framework for training MISR algorithms from bursts of satellite images without requiring high resolution ground truth. This is achieved by adapting the recently proposed frame-to-frame framework to process bursts of satellite images. In addition we propose an architecture based on feature aggregation that allows to fuse a variable number of frames and is capable of handling degenerate samplings while also reducing noise. On synthetic datasets, the proposed self-supervision strategy attains results on par with those obtained with a supervised training. We applied our framework to real SkySat satellite image bursts leading to results that are more resolved and less noisy than the L1B product from Planet.

Related Material


[pdf] [supp]
[bibtex]
@InProceedings{Nguyen_2021_CVPR, author = {Nguyen, Ngoc Long and Anger, Jeremy and Davy, Axel and Arias, Pablo and Facciolo, Gabriele}, title = {Self-Supervised Multi-Image Super-Resolution for Push-Frame Satellite Images}, booktitle = {Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) Workshops}, month = {June}, year = {2021}, pages = {1121-1131} }