Deep Learning Eliminates Massive Dust Storms from Images of Tianwen-1

TL;DR

This paper introduces a deep learning method that leverages Earth-based image dehazing knowledge to effectively remove dust storms from Martian images, significantly enhancing image clarity and topographical detail.

Contribution

It adapts Earth dehazing techniques to Martian imagery by synthesizing dusty images and training a deep model for dust removal, a novel cross-planet application.

Findings

Effective dust storm removal from Martian images

Improved topographical and geomorphological detail

Quantitative and qualitative validation of approach

Abstract

Dust storms may remarkably degrade the imaging quality of Martian orbiters and delay the progress of mapping the global topography and geomorphology. To address this issue, this paper presents an approach that reuses the image dehazing knowledge obtained on Earth to resolve the dust-removal problem on Mars. In this approach, we collect remote-sensing images captured by Tianwen-1 and manually select hundreds of clean and dusty images. Inspired by the haze formation process on Earth, we formulate a similar visual degradation process on clean images and synthesize dusty images sharing a similar feature distribution with realistic dusty images. These realistic clean and synthetic dusty image pairs are used to train a deep model that inherently encodes dust irrelevant features and decodes them into dust-free images. Qualitative and quantitative results show that dust storms can be effectively eliminated by the proposed approach, leading to obviously improved topographical and geomorphological details of Mars.