Change Detection in UWB SAR Images Based on Robust Principal Component Analysis

Christofer Schwartz,Viet T Vu,Leonardo T Duarte,Lucas P Ramos,Mats I Pettersson,Marcelo Da S Pinho,Renato Machado

doi:10.3390/rs12121916

Abstract

This paper addresses the use of a data analysis tool, known as robust principal component analysis (RPCA), in the context of change detection (CD) in ultrawideband (UWB) very high-frequency (VHF) synthetic aperture radar (SAR) images. The method considers image pairs of the same scene acquired at different time instants. The CD method aims to maximize the probability of detection (PD) and minimize the false alarm rate (FAR). Such aim fits into a multiobjective optimization problem, since maximizing the probability of detection generally implies an increase in the number of false alarms. In that sense, varying the RPCA regularization parameter leads to PD variation with respect to FAR, which is known as receiver operating characteristic (ROC) curve. To evaluate the proposed method, the CARABAS-II data set was considered. The experimental results show that RPCA via principal component pursuit (PCP) can provide a good trade-off between PD and FAR. A comparison between the results obtained with the proposed method and a classical CD algorithm based on the likelihood ratio test provides the pros and cons of the proposed method.

Highlights

Among the different approaches in unsupervised separation, robust principal component analysis (RPCA) methods can deal with additive mixing models and are commonly used for foreground detection in video surveillance systems to detect moving objects [1]
The materials and tools used for the method development and evaluation can be summarized as follows: (a) the Coherent All Radio Band System (CARABAS)-II data set of images, (b) an RPCA via principal component pursuit (PCP) algorithm implementation, (c) a classical change detection (CD) algorithm based on the likelihood ratio test
This paper proposed a method based on RPCA via PCP to perform CD in wavelength-resolution synthetic aperture radar (SAR) images

Summary

Introduction

Among the different approaches in unsupervised separation, robust principal component analysis (RPCA) methods can deal with additive mixing models and are commonly used for foreground detection in video surveillance systems to detect moving objects [1]. In [4], RPCA solved via principal component pursuit (PCP) is used in a pre-processing step to decompose the SAR data into two parts, one related to stationary objects and the other to moving targets. By still considering the GOTCHA data set and the RPCA via PCP, [5] presents preliminary results for multi-pass SAR change detection (CD) in X-band. In the context of ground moving target indication (GMTI), [6] proposes the use of RPCA via a relaxed version of PCP (Relaxed PCP) in order to achieve moving target detection in multichannel surveillance radar systems (MC-SAR), presenting computational results using real data from a Chinese X-band

Objectives

Methods

Results

Conclusion