Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Detecting and Refining HiRISE Image Patches Obscured by Atmospheric Dust

Kunal Sunil Kasodekar

TL;DR

The paper tackles automatic filtering of HiRISE images obscured by atmospheric dust, proposing a robust pipeline that couples a fine-tuned ResNet-50 Dust Image Classifier (achieving 94.05% test accuracy) with a denoising system. It compares PCA+SVM, CNN, and deep ResNet baselines for dusty-patch detection, identifying ResNet-50 as the strongest performer and enabling an automated dust-patch storage pipeline. For denoising, it juxtaposes an Upscaled Autoencoder and a Pix2Pix GAN, with Pix2Pix delivering sharper, near-artifact-free reconstructions and SSIM near 1.0 under moderate noise; extreme dust scenarios remain challenging. The work advances automated data cleaning for planetary remote sensing, reducing manual filtering, improving data pipeline efficiency, and enabling post-processing like denoising, with plans for cloud deployment and expanded datasets.

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.

Detecting and Refining HiRISE Image Patches Obscured by Atmospheric Dust

TL;DR

The paper tackles automatic filtering of HiRISE images obscured by atmospheric dust, proposing a robust pipeline that couples a fine-tuned ResNet-50 Dust Image Classifier (achieving 94.05% test accuracy) with a denoising system. It compares PCA+SVM, CNN, and deep ResNet baselines for dusty-patch detection, identifying ResNet-50 as the strongest performer and enabling an automated dust-patch storage pipeline. For denoising, it juxtaposes an Upscaled Autoencoder and a Pix2Pix GAN, with Pix2Pix delivering sharper, near-artifact-free reconstructions and SSIM near 1.0 under moderate noise; extreme dust scenarios remain challenging. The work advances automated data cleaning for planetary remote sensing, reducing manual filtering, improving data pipeline efficiency, and enabling post-processing like denoising, with plans for cloud deployment and expanded datasets.

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.
Paper Structure (14 sections, 8 figures)

This paper contains 14 sections, 8 figures.

Table of Contents

  1. Introduction
  2. Related Work
  3. Dataset
  4. PROPOSED METHODOLOGY
  5. Preprocessing
  6. Classification
  7. Image Noising
  8. Denoising Models
  9. Results
  10. Classifier
  11. Denoiser
  12. Quality of Denoising
  13. Discussion
  14. CONCLUSION

Figures (8)

  • Figure 1: Dusty Patches Obscured by Atmospheric Dust (left), Clear Images of the Martian Landscape (Right) [Resolution: 100x100]
  • Figure 2: Actual Clear Images (left), Denoised Cleaned Images using an Upscaled AutoEncoder Architecture (right) [Resolution: 128x128]
  • Figure 3: Noisy Images for the Corresponding Denoised Images Above [Resolution: 128x128]
  • Figure 4: Upscaled AutoEncoder Architecture
  • Figure 5: Confusion Matrix
  • ...and 3 more figures