利用MATLAB仿真工具比较了自适应滤波器的常用自适应算法LMS算法和RLS算法,论述了基于变步长、正则LMS算法自适应滤波器原理,并针对漏磁检测的缺陷信号和噪声特点,设计了一种适于漏磁信号除噪的自适应滤波器。通过MATLAB程序仿真,试验结果表明该滤波器具有较快的收敛特性和较小的失调量,可满足漏磁检测系统除噪要求。 The MATLAB simulation tools are used to compare LMS algorithm and RLS algorithm which are commonly used in adaptive filter. The principle of adaptive filter based on variable step and regular LMS algorithm is discussed. According to the defect signal and noise characteristic of MFL detection, An Adaptive Filter for Magnetic Flux Leakage Removal. Through the MATLAB program simulation, the experimental results show that the filter has faster convergence and smaller offset, which can meet the requirement of eliminating noise in the magnetic flux leakage detection system.