基于小波包分解的超声相控阵前置阈值降噪研究  

Research on Pre-Threshold Denoising of Ultrasonic Phased Array Based on Wavelet Packet Decomposition

作　　者：杨宸旭 高鸿波[1] 沈佳卉 李秋锋[1] 石文泽[1] 董德秀 陈尧[1] YANG Chenxu;GAO Hongbo;SHEN Jiahui;LI Qiufeng;SHI Wenze;DONG Dexiu;CHEN Yao(Key Laboratory of Nondestructive Testing Technology(MOE),Nanchang Hangkong University,Nanchang Jiangxi 330063,China;Power China Jiangxi Electric Power Engineering Limited Company,Nanchang Jiangxi 330096,China;AECC Shenyang Liming Aero-Engine Co.,LTD.,Shenyang Liaoning 110043,China)

机构地区：[1]无损检测技术教育部重点实验室(南昌航空大学),江西南昌330063 [2]中国电建集团江西省电力设计院有限公司,江西南昌330096 [3]中国航发沈阳黎明航空发动机有限责任公司,辽宁沈阳110043

出　　处：《传感技术学报》2022年第10期1368-1374,共7页Chinese Journal of Sensors and Actuators

基　　金：国家自然科学基金(11764030);江西省主要学科和技术带头人资助计划(20204BCJL22039);江西省自然科学基金重点项目(20212ACB204014)。

摘　　要：超声相控阵由于其独特的优点广泛应用于航空航天工业领域,在实际检测过程中,噪声污染难以避免,仪器电噪声和环境噪声会造成伪像以及成像分辨率低等问题。为此利用小波包分解对超声相控阵检测开展前置阈值降噪研究,首先对多孔碳钢试块进行超声相控阵实验检测,提取B扫、S扫的各个阵元原始数据;然后利用小波包分解检测信号,获得分解信号中最底层各个节点的能量谱;最后提出一种随分解层数而变化的自适应阈值计算方法,并利用改进后阈值函数进行去噪处理,将重构后的信号成像与未降噪信号成像对比。结果表明这种前置阈值降噪方法既可以保留原始信号的特征,也可以对信号中的噪声杂波起到抑制作用;相比带有噪声的B扫、S扫检测信号,降噪后缺陷图像信噪比分别平均提升8.97 dB、8.26 dB,且改进阈值函数降噪后的缺陷图像信噪比相比硬阈值降噪和软阈值降噪,B扫平均提升2.36 dB和2.39 dB;S扫平均提升3.87 dB和3.17 dB。所提方法在工程实际中具有较好的应用前景。Ultrasonic phased arrays are widely used in the aerospace industry due to its outstanding advantages.In the actual inspection process,noise pollution is unavoidable.The electrical noise of the instrument and environmental noise can cause artifacts or low imaging resolution.For this reason,wavelet packet decomposition is used to carry out pre-threshold noise reduction on the ultrasonic phased array detection.Firstly,the ultrasonic phased array experimental detection is performed on the porous carbon steel specimen,and the original data of each element of the B-scan and S-scan are extracted.Secondly,the detection signals are decomposed with wavelet packet to obtain the energy spectrum of each node of the decomposed signal.Finally,an adaptive threshold calculation method that changes with the number of decomposition layers is proposed,the improved threshold function is used for denoising process,and the reconstructed signal imaging is compared with the imaging without performing noise reduction.The results show that the pre-threshold noise reduction method can not only retain the characteristics of the original signal,but also suppress the noise clutter in the signal.Compared with the noisy B-scan and S-scan detection signals,the signal-to-noise ratio of the defect image after denoising is increased by 8.97 dB and 8.26 dB on average,and the signal-to-noise ratio of the defect image after denoising with improved threshold function is 2.36 dB and 2.39 dB higher than those of hard threshold noise reduction and soft threshold noise reduction,and the values of the average improvement of S sweep are 3.87 dB and 3.17 dB compared with those two methods.The proposed method has a better application prospect in engineering practice.

关 键 词：超声相控阵 小波包分解 阈值降噪 聚焦扫查 

分 类 号：TB552[理学—物理]