钢丝绳的检测信号中具有大量的噪声信号，给断丝识别带来很大困难。分析了钢丝绳断丝信号的特征，采用多分辨分析进行消噪，同时讨论了Lipschitz指数，从分解的信号中分别计算出奇异点的Lipschitz指数，并以此作为断丝的特征量进行识别。试验结果表明，该方法可对断丝准确定位，并大大减小了断丝的误判率。

There were lots of noises in the signals of wire rope testing, therefore it was difficult to identify the broken wires in it. The feature of broken wire signals was analyzed, and denoising was carried out by wavelet multi-resolution analysis. Singularity exponent was also discussed, based on which broken wires were identified. Experimental results showed that broken wires could be orientated accurately by the method, and the rate of mistaken detection was greatly decreased.