中文摘要：

煤矿钢芯输送带缺陷电磁检测信号易受到煤矿工况中非平稳强噪声影响,因此,提出一种改进阈值小波的变步长LMS自适应滤波算法对缺陷信号非平稳强噪声进行滤波。该方法首先运用改进阈值小波多尺度分解后的高频信号作为自适应滤波的参考噪声信号,然后运用变步长LMS自适应滤波方法对缺陷信号进行滤波。该方法克服了小波软、硬阈值的缺陷和固定步长LMS自适应滤波中收敛速度和稳态误差对步长因子要求不一致的矛盾。并与改进阈值小波滤波方法进行了对比,结果表明：该方法具有更好的滤波效果,能有效地滤除缺陷信号中的非平稳强噪声。

英文摘要：

Defect electromagnetic testing signals of steel cord conveyor belt in coal mine are vulnerable to be affected by the non-stationary strong noise of working conditions in coal mine. In this paper,a kind of variable step size LMS adaptive filtering algorithm based on improved threshold wavelet was presented,which was used to filter non-stationary noise of defect signals of steel cord conveyor belt. First,in this method,the high frequency signals from improved threshold wavelet multi-scale decomposition were used as reference noise signals of adaptive filtering. Then,the defect signals were filtered by variable step size LMS adaptive filtering method. This method can overcome soft and hard threshold defects of the wavelet,and step size requirement inconsistent contradiction between convergence speed and steady-state error. This method was also compared with improved threshold wavelet filtering method. The results show that the method has better filtering effect,as well as it can effectively filter nonstationary strong noise in defect signals of steel cord conveyor belt.