基于数学形态学的超声C扫描图像校正方法  

Correction Method of Ultrasonic C-scan Image Based on Mathematical Morphology

作　　者：曹宗杰[1] 潘希德[1] 薛锦[1] 王裕文[1] 

机构地区：[1]西安交通大学材料科学与工程学院焊接研究所,西安710049

出　　处：《仪器仪表学报》2005年第z1期681-683,共3页Chinese Journal of Scientific Instrument

摘　　要：通过分析表面波对超声C扫描图像的影响发现,由于超声探头纵向分辨率的限制,对缺陷回波采样过程中叠加的表面回波分量在超声C扫描图像中形成了不均匀分布的图像背景,影响了超声无损评价的准确性。针对上述问题,采用基于数学形态学的形态滤波方法对C扫描图像进行了校正。试验结果表明,通过形态滤波,超声C扫描图像中的不均匀背景得到了校正,同时保留了缺陷的结构特征,进一步对图像进行阈值分割,分割结果准确性明显提高。The influence of surface echo on ultrasonic C-scan image was analyzed firstly. It is found that the surface echo superposed on the defect echo during sampling process, which is due to the limit of axial resolution of ultrasonic probe, results in the non-uniform background in ultrasonic C-scan image. It takes detrimental effect on the accuracy of ultrasonic nondestructive evaluation. To resolve this problem, C-scan image was corrected by the morphological filter based on the mathematical morphology. The experimental results show that not only the non-uniform background in C-scan image was corrected, but also the structure character of defect was reserved after correction. Furthermore, the corrected image was processed by threshold segmentation. The segmented result was more accurate.

关 键 词：超声图像 无损评价 数学形态学 图像处理 

分 类 号：TH7-55[机械工程—仪器科学与技术]