Impact of temporal Doppler on synthetic aperture sonar imagery.

Yan Pailhas,Chris Capus,Samantha Dugelay

doi:10.1121/1.5021250

Abstract

Synthetic Aperture Sonar (SAS) coherently processes the acoustic data acquired along a linear trajectory. The imaging process is in essence an inverse problem where the reflectivity of the seafloor has to be estimated. Several imaging algorithms have been proposed over the years including back-projection algorithms. One commonly assumed hypothesis, however, is that the antenna is motionless during transmission and reception. This hypothesis is known as the start-stop assumption. This paper questions the validity of this assumption, and proposes a full derivation of the SAS processing taking into account the vehicle motion by using the Lorentz transformation. The cell migration for the SAS system is computed and the validity limit of the stop-start assumption depending on the SAS system parameters is derived. Numerical examples of start-stop assumption violations are given and the Doppler cell migration correction on real SAS data are presented and discussed.

Highlights

SAS (Synthetic Aperture Sonar) processing shares the same theoretical principles as SAR (Synthetic Aperture Radar): high resolution imaging is achieved by coherently processing consecutive pings
This paper questions the validity of this assumption, and proposes a full derivation of the SAS processing taking into account the vehicle motion by using the Lorentz transformation
III, as well as the implementation of a modified time domain back-projection (BP) algorithm to compensate for the temporal Doppler

Summary

Introduction

SAS (Synthetic Aperture Sonar) processing shares the same theoretical principles as SAR (Synthetic Aperture Radar): high resolution imaging is achieved by coherently processing consecutive pings. The coherent processing over a large synthetic antenna is feasible if the ping-to-ping antenna displacement is estimated within subwavelength accuracy along the synthetic aperture length. The micronavigation problem for SAS systems has been solved thanks to the use of a multi-element receiver array antenna[1] and the development of the DPCA (Displaced Phase-Centre Antenna) approximation algorithm.[2,3] Unlocking coherent processing has opened a new field of research and the SAS community has grown significantly since. The increase in resolution compared to sidescan sonar systems has made Automatic Target Recognition algorithms more efficient and viable.[4,5,6] As a result, SAS systems are a sensor of choice for large area surveys and Mine Counter-Measures missions especially. The nature of SAS imagery allows coherent processing with various applications including subaperture coherent processing[7,8] or more recently SAS interferometry processing using multiple paths and acquisitions over the same area has been used for incoherent and coherent change detection.[9,10] Various reacquisition patterns have been investigated including circular acquisition, which is a interesting topic as the circular acquisition maximises the aperture of the system and the achievable resolution.[11,12]

Objectives

Results

Conclusion