作者：孙春光1, 何敏1, 曾星星2, 冯肖维1
作者单位：1. 上海海事大学物流工程学院，上海 201306;
2. 华北电力大学控制与计算机工程学院，北京 102206
关键词：电磁层析成像;迭代硬阈值;l1正则化;l2正则化;稀疏成像
摘要：
为提高电磁层析成像(electromagnetic tomography，emt)技术在金属结构缺陷检测时图像重建的效果，研究了基于迭代硬阈值(iterative hard thresholding, iht)算法的稀疏成像方法。该文对传统图像重建算法的出发点、计算过程进行分析，再根据金属结构上缺陷分布的稀疏特性，选择稀疏成像方法；结合l2、l1、l0范数约束下解的特点，选择l0范数进行正则化约束解的范围，采用迭代硬阈值算法进行图像重建，并与landweber迭代算法、tikhonov正则化算法的图像重建效果进行对比。软件仿真和硬件实验均表明：l0范数约束下的迭代硬阈值稀疏成像算法能够提高金属缺陷的图像重建质量；得到的图像相对误差比landweber迭代算法、tikhonov正则化算法降低10%；重建图像所用的时间减少一半。
sparse imaging method of emt metal flaw detection based on iht algorithm
sun chunguang1, he min1, zeng xingxing2, feng xiaowei1
1. college of logistics engineering, shanghai maritime university, shanghai 201306, china;
2. school of control and computer engineering, north china electric power university, beijing 102206, china
abstract: in order to improve the image reconstruction effect of electromagnetic tomography (emt) technology in metal structure defect detection, the sparse imaging method based on iterative hard thresholding (iht) algorithm is studied. this paper analyzes the starting point and calculation process of traditional image reconstruction algorithms, and then selects the sparse imaging method according to the sparse characteristics of the defect distribution on the metal structure. combining the characteristics of solutions under l2, l1, and l0 norm constraints, select l0 norm to regularize the range of constraint solutions, and use iterative hard threshold algorithm for image reconstruction, and compare the image reconstruction effects with landweber iterative algorithm and tikhonov regularization algorithm. both software simulation and hardware experiment show that the iterative hard-threshold sparse imaging algorithm under the constraint of l0 norm can improve the image reconstruction quality of metal defects. the relative error of the obtained image is reduced by 10% compared with landweber iterative algorithm and tikhonov regularization algorithm. the time it takes to reconstruct the image is reduced by half.
keywords: electromagnetic tomography;iterative hard thresholding;l1 regularization;l2 regularization;sparse imaging
2022, 48(2):21-26 收稿日期: 2020-12-10;收到修改稿日期: 2021-01-29
基金项目: 国家自然科学基金资助项目（61503241）
作者简介: 孙春光(1994-),男,河南周口市人,硕士研究生,专业方向为emt金属缺陷的电磁层析成像技术研究
参考文献
[1] dang c, gao j, wang z, et al. multi-step radiographic image enhancement conforming to weld defect segmentation[j]. iet image processing, 2015, 9(11): 943-950
[2] 吴燕, 李洋, 穆海宝, 等. 基于电阻频响法的绕组材质无损检测技术研究[j]. 电测与仪表, 2020, 57(14): 74-82
[3] 许维蓥, 谢烁熳, 洪晓斌. 基于ect的聚合物锂电池组内部电容率检测装置设计[j]. 中国测试, 2018, 44(1): 74-79
[4] 王劭菁, 王丰华, 周东旭, 等. 基于电磁层析成像技术的接地网导体异常检测[j]. 电测与仪表, 2021, 58(6): 140-147
[5] 王琦, 崔莉莎, 汪剑鸣, 等. 基于电磁层析成像的金属缺陷稀疏成像方法[j]. 仪器仪表学报, 2017, 38(9): 2291-2298
[6] 张玲玲, 王化祥, 范文茹. 基于1范数的电阻层析成像图像重建算法[j]. 天津大学学报, 2011, 44(9): 786-790
[7] 王丕涛, 王化祥, 孙犇渊. 基于l_1范数的电容层析成像图像重建算法[j]. 中国电机工程学报, 2015, 35(18): 4709-4714
[8] ye j, wang h, yang w. image reconstruction for electrical capacitance tomography based on sparse representation[j]. ieee transactions on instrumentation and measurement, 2015, 64(1): 89-102
[9] chartrand r, yin w. iteratively reweighted algorithms for compressive sensing[c]//2008 ieee international conference on acoustics, speech and signal processing, 2008.4518498.
[10] 张立峰, 朱炎峰. 极限学习机在电容层析成像中的应用[j]. 电测与仪表, 2020, 57(9): 146-152
[11] 肖理庆, 王化祥, 厉丹. 改进landweber预迭代ert图像重建算法[j]. 中国电机工程学报, 2013, 33(23): 118-125
[12] xu f, deshpande m. iterative nonlinear tikhonov algorithm with constraints for electromagnetic tomography[j]. ieee journal of selected topics in applied earth observations and remote sensing, 2012, 5(3): 707-716
[13] 荣光李, 黄尉. 基于子空间追踪算法的稀疏子空间聚类[j]. 合肥工业大学学报(自然科学版), 2019, 42(7): 999-1004
[14] polat a, matela n, mota a m, et al. implementation of majorization-minimization (mm) algorithm for 3d total variation minimization in dbt image reconstruction[c]//2016 ieee nuclear science symposium, medical imaging conference and room-temperature semiconductor detector workshop (nss/mic/rtsd), 2016.
[15] blumensath t, davies m e. iterative hard thresholding for compressed sensing[j]. applied and computational harmonic analysis, 2009, 27(3): 265-274
[16] blumensath t, davies m e. iterative thresholding for sparse approximations[j]. journal of fourier analysis and applications, 2008, 14(5): 629-654
[17] 曾凡凡. 基于l_p范数的压缩感知和深度学习磁共振图像重建研究[d]. 合肥: 中国科学技术大学, 2020.
