Adaptive Sonar Imaging Using Aperture Coherence

A class of computationally efficient adaptive imaging methods originating from the coherence factor (CF) has been proposed for improved medical ultrasound imaging. These methods are based on the idea that when steering the receiver toward a point of interest, the backscattered energy from this point exhibits a high degree of aperture coherence, while random noise, multipath interference, and sidelobe energy do not. Aperture coherence can be understood as a normalized measure of the degree of signal variability across the receiver array. This paper presents a study of the use of aperture-coherence-based methods for improved sonar imaging, with particular emphasis on a recently introduced robust implementation known as the scaled Wiener postfilter (SWiP). We show that while the CF has strong noise suppression capabilities and performs well on point targets, it lacks robustness in low signal-to-noise ratio (SNR) scenarios and introduces undesirable artifacts in speckle scenes. The SWiP is closely related to the CF, but contains a single-user-defined parameter, which allows the method to be tuned to suit the application needs. The SWiP can be tuned to offer robustness in a speckle environment such as when imaging the seafloor, or for strong noise suppression capabilities. This makes it a promising method for a wide range of sonar applications. We base our conclusions on simulated data from a constructed speckle scene as well as experimental data from the SX90 fish-finding sonar and sidescan data from the HISAS 1030 sonar. Our results show that the SWiP offers improved edge and shadow definitions and reduced sidelobe levels when compared to the conventional delay-and-sum (DAS) beamformer. These improvements do not compromise the image resolution, and they come at a low computational cost.