基于相似波形的漏磁检测缺陷信号小波分解及提取方法

Wavelet decomposition and extraction method of defect magnetic flux leakage testing signals based on waveform similarity

漏磁检测技术被广泛应用于铁磁构件缺陷的检测和量化评估,其对缺陷形状、尺寸的三维反演成像、量化能力,尤其是对常见的自然(腐蚀、疲劳)复杂缺陷的三维成像能力,是漏磁检测技术水平的核心标志之一。由于实际缺陷形状不规则、常包含多个坑点,致使其漏磁检测信号彼此影响。为提升复杂缺陷的成像精度和数据处理速度,从漏磁信号里自动分解并剥离出对应缺陷的信号是很重要的前置工作。本文提出一种基于相似波形的漏磁检测缺陷信号小波分解及提取方法,用于自动分解并拾取每个缺陷的漏磁检测信号,并自动确定每个缺陷的位置及边界,以备后续利用其完成对缺陷的三维反演成像。该方法基于实际检测到的缺陷漏磁信号的形态特征,匹配与其波形最相似的小波基;用选取的小波基对每个通道的漏磁检测信号进行多尺度分解,得到每个通道信号的小波高频系数;再根据小波系数特征确定每个通道信号中存在的缺陷数量和缺陷所在的位置,并提取每个通道各处缺陷的漏磁信号;最终通过判定缺陷信号相邻通道的连通性,将所在连通区域内的缺陷信号进行集成,得到各处缺陷的信息集,从而实现对所有缺陷漏磁检测信号的提取。本文提出的方法可用于对复杂缺陷漏磁检测信号的准确和快速提取、分离和集成,为后续缺陷的三维反演成像奠定基础。

Magnetic flux leakage testing(MFLT)technology has been widely used in engineering to identify and measure defects in ferromagnetic components.Its capability in reconstructing defects in the three-dimensional(3D)form,especially for general complex defects formed naturally(such as by corrosion and/or fatigue),is one of the primary indexes representing the technological advances.Because the actual defect shape is irregular and often contains multiple pits,its magnetic flux leakage detection signals affect each other.To improve the accuracy of reconstructing defects in 3D and speed up the data processing,the prerequisite is to automatically strip off and extract the valid defect data from the MFLT signals.To effectively locate,separate,and integrate information for reconstructing each defect,a wavelet decomposition and extraction method is proposed.Based on the morphological characteristics of the MFLT signals,a wavelet basis is selected that matches the characteristics of the defect signals closely.This wavelet basis is used in the multi-scale decomposing of the MFLT signals and in the calculating of the associated wavelet high frequency coefficients.Then the wavelet coefficient is utilized in identifying and locating the defects in each channel of the signals,as well as in extracting the MFLT data corresponding to each defect.Finally,by examining the signals of each defect in a channel and integrating all the associated data in the adjacent channels,a complete set of data related to each defect is obtained.The proposed method can be applied to quickly identify,extract and integrate the MFLT data for all defects,which provides the basis for the 3D defect reconstruction.