Abstract:The use of Synthetic Aperture Radar (SAR) has greatly advanced our capacity for comprehensive Earth monitoring, providing detailed insights into terrestrial surface use and cover regardless of weather conditions, and at any time of day or night. However, SAR imagery quality is often compromised by speckle, a granular disturbance that poses challenges in producing accurate results without suitable data processing. In this context, the present paper explores the cutting-edge application of Quantum Machine Learning (QML) in speckle filtering, harnessing quantum algorithms to address computational complexities. We introduce here QSpeckleFilter, a novel QML model for SAR speckle filtering. The proposed method compared to a previous work from the same authors showcases its superior performance in terms of Peak Signal-to-Noise Ratio (PSNR) and Structural Similarity Index Measure (SSIM) on a testing dataset, and it opens new avenues for Earth Observation (EO) applications.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is: how to effectively remove speckle in Synthetic Aperture Radar (SAR) images by using Quantum Machine Learning (QML) techniques, thereby improving the quality and reliability of SAR images. ### Problem Background Synthetic Aperture Radar (SAR) images play an important role in earth observation and can provide detailed surface information under any weather and day - night conditions. However, the quality of SAR images is often interfered by speckle noise, which appears as granular perturbations in the images and seriously affects the interpretation and application effects of the images. Therefore, removing speckle noise is crucial for obtaining reliable and operable earth observation data. ### Core Contributions of the Paper To meet this challenge, the paper introduces a new quantum machine - learning model - **QSpeckleFilter**, which combines quanvolution and traditional speckle denoisers. Specifically: 1. **Quantum Convolution Pre - processing**: - Use quantum convolution operations to pre - process the input data, expanding the original single - channel data into a multi - channel feature map, enhancing the model's ability to extract features. - The pre - processed data is used to train an improved version of the speckle denoiser. 2. **Improved Speckle Denoiser Architecture**: - Use the feature map generated by quantum convolution to estimate the speckle noise in the image and remove the noise from the original SAR image through skip connections and subtraction layers. - This method avoids the logarithmic scaling and division operations in traditional methods, reducing computational errors and preserving the statistical characteristics of the data. 3. **Performance Evaluation**: - On the test data set, the QSpeckleFilter model is significantly superior to previous works in terms of metrics such as Peak Signal - to - Noise Ratio (PSNR) and Structural Similarity Index (SSIM), with improvements of 13% and 8% respectively. ### Conclusion The paper demonstrates the great potential of quantum machine learning in SAR speckle filtering and proves that quantum convolution, as a powerful pre - processing technique, can significantly improve the effect of speckle denoising. This not only provides higher - quality SAR images for earth observation applications but also opens up new directions for future research.

QSpeckleFilter: a Quantum Machine Learning approach for SAR speckle filtering

Unassisted Quantitative Evaluation Of Despeckling Filters

Speckle Noise Analysis for Synthetic Aperture Radar (SAR) Space Data

SAR Despeckling Using Multiobjective Neural Network Trained with Generic Statistical Samples.

A New Nonlocal Iterative Trilateral Filter for SAR Images Despeckling

A new SAR image filter for preserving speckle statistical distribution

SAR Speckle Removal Using Hybrid Frequency Modulations

A Novel Method of Speckle Reduction and Enhancement for SAR Image.

SAR Image Despeckling Based on Non-local Means Filter

Deep Learning Based Speckle Filtering for Polarimetric SAR Images. Application to Sentinel-1

SAR2SAR: a semi-supervised despeckling algorithm for SAR images

A Novel SAR Image Despeckling Method Based on Local Filter With Nonlocal Preprocessing

A Trilateral Filter for Video SAR Speckle Noise Reduction

Speckle Reduction in Polarimetric SAR Imagery with Stochastic Distances and Nonlocal Means

Quantum Machine Learning for Remote Sensing: Exploring potential and challenges

Non-Local SAR Image Despeckling Based on Sparse Representation

An Advanced SAR Image Despeckling Method by Bernoulli-Sampling-Based Self-Supervised Deep Learning

SAR Image Despeckling Based on Denoising Diffusion Probabilistic Model and Swin Transformer

Speckle2Void: Deep Self-Supervised SAR Despeckling with Blind-Spot Convolutional Neural Networks

SAR Despeckling using a Denoising Diffusion Probabilistic Model

Adaptive Speckle Filter for Multi-Temporal PolSAR Image with Multi-Dimensional Information Fusion