论文部分内容阅读
针对目前常见的稀疏字典缺乏针对性,在合成孔径医学超声成像中的应用效果不佳,难以在低压缩率下保证重构图像质量的问题,本文设计了一种高效能的稀疏字典。根据超声回波信号是由发射脉冲信号经过不同延时衰减后叠加的特点,利用发射脉冲作为基函数构造稀疏字典,回波信号在该稀疏字典确定的变换域中具备很好的稀疏性,理论上能使其稀疏表示系数的稀疏度等于超声阵元接收到的反射回波数。通过FieldⅡ对简单点目标和复杂目标的仿真结果表明:在相同的重构算法和压缩率下该稀疏字典重构的平均绝对误差明显小于常见的稀疏字典,其值仅为DWT的几分之一,DFT和DCT的几十分之一,能让回波信号以更低的压缩率实现相同的恢复效果。本文最后使用体模的实际采集数据对算法的实际效果进行检测,实验结果也与仿真结果基本一致。基于该稀疏字典的压缩感知算法可以进一步减少合成孔径成像所需存储的数据量、降低系统的复杂度。
In view of the lack of pertinence of common sparse dictionaries, the application in Synthetic Aperture Medical Ultrasound is not effective and it is difficult to guarantee the quality of reconstructed image with low compression rate. A high-performance sparse dictionary is designed in this paper. According to the characteristic that the ultrasonic echo signal is superposed after the transmitted pulse signal is attenuated by different delays, the sparse dictionary is constructed by using the emission pulse as a basis function, and the echo signal has good sparsity in the transform domain determined by the sparse dictionary. The theory The sparsity that makes it sparse is equivalent to the number of reflected echoes received by the ultrasound array element. The simulation results of Field Ⅱ on simple point targets and complex targets show that the average absolute error of sparse dictionary reconstruction under the same reconstruction algorithm and compression ratio is significantly less than that of common sparse dictionaries, which is only a fraction of DWT A few tenths of the DFT and DCT allows the echo signal to achieve the same recovery at a lower compression rate. In the end, we use the actual data collected from the phantom to test the actual effect of the algorithm. The experimental results are also consistent with the simulation results. Compressive sensing algorithm based on this sparse dictionary can further reduce the amount of data stored in synthetic aperture imaging and reduce the complexity of the system.