Page 90 - 《应用声学》2023年第3期

528 2023 年 5 月

表 1 不同采样率下 MSBL 算法和 SBL 算法 Bai Zhiliang, Chen Shili, Jia Lecheng, et al. Compressed
重构耗时 sensing of ultrasonic phased array signal in turbine disc
rims inspection based on greedy algorithms[J]. Journal of
Table 1 Cost time of MSBL and SBL al-
Mechanical Engineering, 2018, 54(18): 33–41.
gorithms with different SRs
[6] Holmes C, Drinkwater B W, Wilcox P D. Post-processing
(单位：s) of the full matrix of ultrasonic transmit-receive array data
for non-destructive evaluation[J]. NDT & E International,
采样率/% 5 10 15 20 25 30
2005, 38(8): 701–711.
MSBL 222 85 47 32 25 22
[7] 郭俊锋, 王茁. 一种基于多测量向量模型的机械振动信号联合
SBL 12131 7542 5664 4662 4258 3915
稀疏重构方法 [J]. 振动与冲击, 2021, 40(1): 254–263.
Guo Junfeng, Wang Zhuo. A joint sparse reconstruction
3 结论 method for mechanical vibration signals based on multi-
measurement vector model[J]. Journal of Vibration and
(1) 本文将 MMV 模型用于超声 FMC 数据重 Shock, 2021, 40(1): 254–263.
构和 TFM 成像，解决了 SMV 模型存在重构精度不 [8] Berg E V D, Friedlander M P. Theoretical and empir-
ical results for recovery from multiple measurements[J].
足和效率低的问题。
IEEE Transactions on Information Theory, 2010, 56(5):
(2) 实验比较不同采样率下 MSBL算法和 SBL 2516–2527.
算法的重构效果。在 100 MHz 采样频率下，MSBL [9] Cotter S F, Rao B D, Kjersti E, et al. Sparse solutions to
算法使用 15% 的采样点进行重构，归一化均方误差 linear inverse problems with multiple measurement vec-
tors[J]. IEEE Transactions on Signal Processing, 2005,
仅为 2%；相同误差下，SBL 算法约需 25% 采样点， 53(7): 2477–2488.
且重构耗时比 MSBL 算法高两个数量级，反映了 [10] Eldar Y C, Rauhut H. Average case analysis of mul-
MMV模型在FMC数据中应用的优越性。 tichannel sparse recovery using convex relaxation[J].
(3) MMV 模型的重构性能优于 SMV 模型，但 IEEE Transactions on Information Theory, 2009, 56(1):
505–519.
重构原 FMC 数据的过程增加了 TFM 实时成像的 [11] 李卫卫. 基于信号结构特性的压缩感知关键技术研究 [D]. 北
复杂度。融合压缩感知与波束成形方法是未来的研 京邮电大学, 2015.
究方向与目标。 [12] Wipf D P, Rao B D. An empirical Bayesian strategy
for solving the simultaneous sparse approximation prob-
lem[J]. IEEE Transactions on Signal Processing, 2007,
参 考 文 献 55(7): 3704–3716.
[13] 徐建平, 皮亦鸣, 曹宗杰. 基于贝叶斯压缩感知的合成
[1] 张鑫, 林莉, 金士杰. 奥氏体不锈钢窄间隙焊缝侧壁未熔合相 孔径雷达高分辨成像 [J]. 电子与信息学报, 2011, 33(12):
控阵超声检测 [J]. 应用声学, 2021, 40(1): 97–102. 2863–2868.
Zhang Xin, Lin Li, Jin Shijie. Phased array ultrasonic Xu Jianping, Pi Yiming, Cao Zongjie. SAR imaging based
testing for lack of sidewall fusion in narrow gap welding of on Bayesian compressive sensing[J]. Journal of electronics
austenitic stainless steel[J]. Journal of Applied Acoustics, & information technology, 2011, 33(12): 2863–2868.
2021, 40(1): 97–102. [14] Zhang M, Markovsky I, Schretter C, et al. Compressed ul-
[2] 杨平, 施克仁. 相控阵超声二维稀疏阵列的优化设计应用研 trasound signal reconstruction using a low-rank and joint-
究 [J]. 应用声学, 2008, 27(2): 148–154. sparse representation model[J]. IEEE Transactions on Ul-
Yang Ping, Shi Keren. A novel method to design 2D trasonics, Ferroelectrics, and Frequency Control, 2019,
sparse array for ultrasonic phased array[J]. Journal of Ap- 66(7): 1232–1245.
plied Acoustics, 2008, 27(2): 148–154. [15] 王法松, 张林让, 周宇. 压缩感知的多重测量向量模型与算法
[3] Donoho D L. Compressed sensing[J]. IEEE Transactions 分析 [J]. 信号处理, 2012, 28(6): 785–792.
on Information Theory, 2006, 52(4): 1289–1306. Wang Fasong, Zhang Linrang, Zhou Yu. Multiple mea-
[4] 杨晓霞, 薛彬, 靳世久, 等. 压缩传感在超声相控阵检测 surement vectors for compressed sensing: model and
系统中的应用研究 [J]. 电子测量与仪器学报, 2015, 29(9): algorithms analysis[J]. Signal Processing, 2012, 28(6):
1286–1294. 785–792.
Yang Xiaoxia, Xue Bin, Jin Shijiu, et al. Application [16] Liebgott H, Prost R. Pre-beamformed RF signal recon-
study on compressed sensing in ultrasonic phased array struction in medical ultrasound using compressive sens-
detection system[J]. Journal of Electronic Measurement ing[J]. Ultrasonics, 2013, 53(2): 525–533.
and Instrumentation, 2015, 29(9): 1286–1294. [17] Jing L, He Q, Luo J. A compressed sensing strategy for
[5] 白志亮, 陈世利, 贾乐成, 等. 基于贪婪算法的汽轮机叶轮相 synthetic transmit aperture ultrasound imaging[J]. IEEE
控阵信号压缩感知 [J]. 机械工程学报, 2018, 54(18): 33–41. Transactions on Medical Imaging, 2017, 36(4): 878–891.