Compressive Sensing in Acoustic Imaging
Abstract
Acoustic sensing is at the heart of many applications, ranging from underwater sonar and nondestructive testing, to the analysis of noise and their sources, medical imaging, and musical recording. This chapter discusses a palette of acoustic imaging scenarios where sparse regularization can be leveraged to design compressive acoustic imaging techniques. Nearfield acoustic holography (NAH) serves as a guideline to describe the general approach. By coupling the physics of vibrations and that of wave propagation in the air, NAH can be expressed as an inverse problem with a sparsity prior, and addressed through sparse regularization. In turn, this can be coupled with ideas from compressive sensing to design semi-random microphone antennas, leading to improved hardware simplicity, but also to new challenges in terms of sensitivity to a precise calibration of the hardware and software scalability. Beyond NAH, this chapter shows how compressed sensing is being applied to other acoustic scenarios such as active sonar, sampling of plenacoustic function, medical ultrasound imaging, localization of directive sources and interpolation of plate vibration response.