Coded excitation reconstruction by impulse response estimation and retrospective acquisition

Abstract

Coded excitation (CE) imaging enables large time-bandwidth waveform transmissions, which are often pulse-compressed by matched filtering (MF). Much research investigates schemes for reducing MF sidelobes to decrease clutter with distributed targets. The goal of this presentation is to harmonize CE transmission with the image quality of Synthetic Transmit Aperture (STA) imaging at high frame rates. We propose a two-step process that estimates medium impulse responses (IR), then retrospectively images using the IR set. In this way, the probing code sequence is used optimally yet is decoupled from the imaging step, circumventing the sidelobe problem. The method first estimates transmit-receive (TR) pair IR of the acoustic medium, using random codes transmitted simultaneously across channels and acquisition intervals. Linear model theory solves for unbiased estimates of the IR set. In the second step, retrospective STA imaging excites the estimated IR set, using it as a simulation of the medium. Retrospective transmit signals and apodizing can differ at each pixel. We show the resulting beamformed pixels are unbiased estimates of STA pixels, and that time diversity in codes improves estimation error over static codes. To address tissue motion, the model is extended to polynomial and Fourier basis. We also demonstrate retrospective transmission of a transducer-compensated pulse in the imaging step, bypassing the limited precision of tri-state transmitter hardware.