图像合成在超声兰姆波层析成像中的应用  

Application of Image Compounding in Ultrasonic Lamb Wave Tomography

作　　者：张海燕[1] 刘倩玮[1] 范仕轩[1] 吕东辉[1] 

机构地区：[1]上海大学通信与信息工程学院,上海200072

出　　处：《上海大学学报（自然科学版）》2007年第4期430-433,共4页Journal of Shanghai University:Natural Science Edition

基　　金：国家自然科学基金资助项目(10504020)

摘　　要：在兰姆波层析成像中,由于数据采集系统结构的局限性,致使所得数据不完全,从而引起重建图像的质量下降.图像合成是一种重要的增强图像信息的方法.将图像合成技术引入兰姆波层析成像中,在不同换能器布置方式下对含缺陷板材进行采样,并利用联合迭代重建(SIRT)算法重建缺陷,然后使用均值合成、平方根均值合成和几何均值合成法分别对所获得的图像进行合成.实验结果显示,合成之后的图像视觉上更接近于缺陷的真实形状,而通过对图像质量参数的定量化分析进一步证实了合成之后的图像质量得到显著改善.In Lamb wave tomography, the limitation caused by data acquisition geometry generally results in incomplete projection data, therefore poor reconstruction quality. Image compounding is an effective method to improve the image quality. We apply image compounding in Lamb wave tomography. Incomplete data are acquired on an aluminum plate sample with machined defect using different transducers configurations, and images reconstructed using simultaneous iterative reconstruction technique （SIRT）. Three different methods of image compounding are used, which are average compounding, root-mean-square compounding and geometric mean compounding. The results show that the imaged flaw is visually closer to its real shape, and the reconstruction quality significantly improved after compounding.

关 键 词：兰姆波 层析成像 图像合成 

分 类 号：TG115.28[金属学及工艺—物理冶金]