Denoising Synthetic Aperture Radar / Aerial Images Using HOTV Deep Learning Models with Bayesian MAP Approach

Ashok Hake, Krishnendu Remesh, Vishal Subhash Chavan

2025

Abstract

Denoising plays an essential role in Synthetic Aperture Radar (SAR) and aerial image restoration. These images are distorted with various noises due to atmospheric changes. Therefore, the images should be analyzed using proper restoration and enhancement techniques. Many authors proposed traditional and deep learning models to perform this task. This paper employed the Bayesian Maximum A Posteriori (MAP) approach to the Higher Order Total Variation (HOTV) deep learning model. We assumed that the Poisson noise distorts the images. We also used the model to restore the images degraded by noises such as Gamma, Gaussian, and Rayleigh. Quantitative and qualitative analyses are provided.

Paper Citation

in Harvard Style

Hake A., Remesh K. and Chavan V. (2025). Denoising Synthetic Aperture Radar / Aerial Images Using HOTV Deep Learning Models with Bayesian MAP Approach. In Proceedings of the 3rd International Conference on Futuristic Technology - Volume 2: INCOFT; ISBN 978-989-758-763-4, SciTePress, pages 662-668. DOI: 10.5220/0013599700004664

in Bibtex Style

@conference{incoft25,
author={Ashok Hake and Krishnendu Remesh and Vishal Chavan},
title={Denoising Synthetic Aperture Radar / Aerial Images Using HOTV Deep Learning Models with Bayesian MAP Approach},
booktitle={Proceedings of the 3rd International Conference on Futuristic Technology - Volume 2: INCOFT},
year={2025},
pages={662-668},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013599700004664},
isbn={978-989-758-763-4},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 3rd International Conference on Futuristic Technology - Volume 2: INCOFT
TI - Denoising Synthetic Aperture Radar / Aerial Images Using HOTV Deep Learning Models with Bayesian MAP Approach
SN - 978-989-758-763-4
AU - Hake A.
AU - Remesh K.
AU - Chavan V.
PY - 2025
SP - 662
EP - 668
DO - 10.5220/0013599700004664
PB - SciTePress