Pseudo-Zernike Moments Based Sparse Representations for SAR Image Classification

TL;DR

This paper introduces a novel radar image classification method using pseudo-Zernike moments to enhance sparse representations, leveraging invariance properties and complex signatures, validated on the MSTAR dataset.

Contribution

It presents a new approach combining pseudo-Zernike moments with sparse representations for improved SAR image classification.

Findings

Effective classification on MSTAR dataset

Enhanced dictionary redundancy with auxiliary atoms

Utilization of invariance properties improves robustness

Abstract

We propose radar image classification via pseudo-Zernike moments based sparse representations. We exploit invariance properties of pseudo-Zernike moments to augment redundancy in the sparsity representative dictionary by introducing auxiliary atoms. We employ complex radar signatures. We prove the validity of our proposed methods on the publicly available MSTAR dataset.