基于压缩感知理论的高分辨率超声成像技术研究

Recently, a lot of problems are in urgent need to solve in high resolution ultrasound imaging system, such as its high sampling rate, large sampling data valume, the system complex and its high cost. Based on this situation, we take advantage of the compressive sensing theory to realize high resolution ultrasound imaging, which provides a new train of thought to solve these problems. In the compressive sensing theory, sampling and compression of the signal are proceeded at the same time for signal sparseness, so we can precise reconstruct the target signal from less sampling data. It has been demonstrated that the required data volume, the complexity and the cost of the ultrasound imaging system were substantially reduced by using the compressive sensing theory. In this project, we will firstly study the sparse representation of signals, the observation matrix and reconstruction algorithm. Secondly, we will theoretical study the principle of ultrasonic imaging based on the compressive sensing theory in accordance with the characteristics of the ultrasonic imaging. By studying the spatial modulation technology, focusing, and single point detection technology of the ultrasonic signal, we can master the technique and build up the ultrasonic imaging system based on the compressive sensing theory. Thirdly, we will analysis of the key factors influencing the resolution of ultrasound imaging based on the compressive sensing theory. Finally, the reconstruction algorithm will be optimized for the study of high resolution rapid ultrasonic imaging technology. These studies will provide scientific theory and experimental basis for solving the problems of high resolution ultrasound imaging technology.

高分辨率超声成像系统中面临的采样率高、数据量大、系统复杂、成本高等问题，是当前一项迫切需要解决的科技问题，利用压缩感知理论实现高分辨率超声成像为解决这个问题提供了一个新的思路。压缩感知理论利用信号的稀疏性，将信号采样和压缩同时进行，利用少量的采样数据实现目标信号的精确重构。研究证实，利用压缩感知理论可以减少超声成像所需的数据量、降低系统成本和复杂程度。本项目将从信号的稀疏表示、观测矩阵和重构算法出发，根据超声成像特点，开展深入的理论研究，揭示基于压缩感知理论的超声成像机理。通过对超声信号空间调制技术、超声信号聚焦技术和超声信号单点探测技术的研究，掌握基于压缩感知理论的超声成像实现技术，搭建成像系统。分析影响基于压缩感知理论的超声成像分辨率的关键因素，优化重构算法，开展高分辨率快速超声成像技术的研究。期望通过该研究，为解决高分辨超声成像技术所遇到的问题提供科学理论和实验基础。

超声成像技术利用超声信号穿透不透光的物体，获取物体内部结构的声学特性信息，广泛应用于无损检测和医学成像领域。随着超声成像技术的发展，人们对超声成像质量要求越来越高，这就使得在成像过程中需要采集和存储的数据量越来越大，系统也越来越复杂。在本项目中，我们将压缩感知理论与超声成像技术相结合，利用少量的采样数据实现了超声高分辨率成像。首先我们对压缩感知理论和超声成像机理进行了大量的理论研究，提出了基于压缩感知的超声高分辨单像素成像模型和基于压缩感知的超声衍射多像素超分辨成像模型，实现了压缩感知理论和超声成像技术的有机结合。基于理论研究成果，我们搭建了基于压缩感知理论的非超分辨超声单像素成像系统和超声关联超分辨单像素成像系统。我们设计制作了超声发射阵列空间调制驱动电路，通过对多路高频小信号控制和功率放大，实现了超声发射阵列的超声信号空间调制，并利用单探测器实现超声信号的全局信息采集，通过优化后的恢复算法，实现了超声高分辨率成像。另外我们将压缩感知成像理论的研究结果与傅里叶层叠衍射成像理论相结合，提出了一种使用单相素探测空间频谱的相位成像方法。我们还将压缩感知理论应用到高光谱成像中，提出了基于单像素技术的傅里叶变换光谱成像技术，搭建了高光谱成像装置。该项目的执行，为我们进一步开展压缩感知和超声成像方面的研究奠定了基础。