一种漏磁信号不变特征提取方法

A Method for Obtaining Invariant Feature of Magnetic Flux Leakage Signals

为了提高漏磁缺陷的重构精度,提出一种基于径向基函数(RBF)的漏磁信号不变特征提取的方法。该方法在对漏磁检测过程中速度影响进行全面分析的基础上,建立二维有限元模型,采集不同速度下的漏磁信号,将速度影响下的漏磁信号作为RBF网络的输入,静态下的漏磁信号作为输出进行训练。实验结果表明:这种方法实现了速度效应的补偿,提取出了漏磁信号的不变特征,而且具有较高的精度,是一种有效的方法。

To improve the defect reconstruction precision of magnetic leakage flux, a method to extract the invariant feature of leakage flux signals was proposed based on radial basis function（RBF） , analyzed the velocity＇s influence in the leakage flux inspection, and established 2-D finite element model, the method acquires the leakage flux signals in different velocities, takes the signals with velocity influence as the input of RBF network, and the static leakage signal as the output, to train the network. The experiment results show that it can compensate the velocity effect and extract the invariant feature of leakage flux signal precisely and effectively.