Microwave-Induced Thermoacoustic Compressive Imaging With Metamaterial Coding

The microwave-induced thermoacoustic (TA) sensing has been proved to promise great potential in clinical and biomedical applications, like breast cancer detection and therapy. Recently, this technology has been exploited to boost its use in subsurface geophysical environments, to image the fluid flow transport in porous media. TA imaging in the geological field signifies a more simplified system with the features of large-scale, fast-scanning and real-time monitoring. As for conventional TA sensing system, it entails a large number of measurements for reconstructing a high-resolution image, which inevitably requires one single ultrasound transducer or an array of transducers in conjunction with a data acquisition system equipped with one or multiple channels. Or one shot is taken with deploying a vast of transducers in a stationary data acquisition system. This leads to long scan time and heavy computational burden for the subsequent image reconstruction. All of them leads to either a considerably slow or a dramatically costly TA system. It greatly limits the potential application of TA technology to imaging fluid flow in a geological media, being too slow to capture the flow front or inducing the excessive image smearing. The pressing need in this work is to reduce the highly intensive sampling of TA measurements without jeopardizing imaging resolution or sensing capacity.

Hitherto, researchers are seeking to address these limitations, and one remarkable solution among them is proposed to develop compressive sensing (CS)-based TA sensing. In the field of TA, the effectiveness of CS essentially depends on the sparsity of the signals emitting from the object. For this reason, the CS can realize a high-quality reconstructed image when the unknown signal possesses a good sparsity. Aiming at favoring this demand by CS, the artificial meta-material (MM) resonators are studied in this work to effectively randomize the propagated TA signals originating from the targets. Inspired by this, the reflection and transmission coefficients of analytical MM resonators are examined in the frequency regime, to ensure the applied MM resonators having a positive effect on the randomness of incident wave patterns. Next, the computational forward model is established to derive the entire time-domain TA waves in simulation: the TA waves are originated from the target under microwave irradiation, then randomized when passing through MM resonator array, and finally received by receivers. Additionally, the inverse model is built upon the background theory to obtain the inversed TA wave fields excited by the source where it was receivers. And the image reconstruction is yield by linearly solving the Helmholtz equation with the consensus-based Alternative Direction Method of Multiplier (ADMM) algorithm. A parametric study of different cell size of MM-array is performed to optimize the sensing capacity. 

Prior to the author’s knowledge, the research endeavor of CS-based TA system utilizing the MM resonator array is rare and has much work to make the practical implementation possible. To bridge the gap, this paper presents and assesses proof-of-concept imaging results of proposed MM-coded TA system and a regular TA system without any presence of Metamaterials, respectively. Based on the comparison under the same scenario, the MM-coded CS-TA system is shown to achieve a higher resolution of an image reconstructed. To yield a similar image, this new system needs less than half the number of measurements as required by a conventional system. The work paves the way for enabling the TA technology for the use of real-time monitoring in geological fields.