【摘要】 研究了在超声检测时利用独立分量分析法对检测信号进行噪声处理技术,在试验中采用Matlab软件对人工加工的焊缝缺陷检测出来的源信号数据进行独立分量分析迭代分离后,再将所得的信号数据分离出源信号与噪声信号,经处理后恢复的缺陷图像明显可以改善之后的焊缝质量评定准确度.结果表明,采用独立分量分析法对噪声分离后更能精确识别检测信号中的缺陷信息,减少由于源信号与噪声的重叠造成了对焊缝缺陷的误判.更多还原

【Abstract】 This paper studied the signal porcess technology by independent component analysis during ultrasonic testing.It also adapted the software of Matlab to analyze and divide the independent component which is examined from the artificial examination of the defects of the welding,and divided the original signals and the noise signals out of the gained signal data.It could obviously improve the evaluation of the quality of the welding.The result showed that the data of the experiment could be more correct,at the same time,the defects and the missing of the welding caused by the noise could be better prevented.更多还原