Detecting and Refining HiRISE Image Patches Obscured by Atmospheric Dust

TL;DR

This paper presents an automated system for detecting and enhancing dusty Mars surface images captured by HiRISE, improving data collection efficiency during dust storms using deep learning classifiers and image restoration techniques.

Contribution

It introduces a Dust Image Classifier fine-tuned on Resnet-50 with high accuracy and a pipeline for automatic classification and denoising of obstructed images using Auto Encoder and Pix2Pix GAN.

Findings

Dust Image Classifier achieves 94.05% accuracy.

Auto Encoder denoises images with 0.75 SSIM.

Pix2Pix GAN enhances image quality with 0.99 SSIM.

Abstract

HiRISE (High-Resolution Imaging Science Experiment) is a camera onboard the Mars Reconnaissance orbiter responsible for photographing vast areas of the Martian surface in unprecedented detail. It can capture millions of incredible closeup images in minutes. However, Mars suffers from frequent regional and local dust storms hampering this data-collection process, and pipeline, resulting in loss of effort and crucial flight time. Removing these images manually requires a large amount of manpower. I filter out these images obstructed by atmospheric dust automatically by using a Dust Image Classifier fine-tuned on Resnet-50 with an accuracy of 94.05%. To further facilitate the seamless filtering of Images I design a prediction pipeline that classifies and stores these dusty patches. I also denoise partially obstructed images using an Auto Encoder-based denoiser and Pix2Pix GAN with 0.75 and 0.99 SSIM Index respectively.