采用自主研发的激光声磁检测系统对钢轨疲劳损伤进行了无损检测研究。选择含有不同尺寸的表面裂纹和表面竖直孔的人工伤钢轨作为检测对象，进行了测试分析。结果表明,缺陷信号的幅度随缺陷深度的增加而增大。为了抑制检测过程中噪声对检测结果的影响，采用Matlab设计软阈值小波去噪算法并对信号进行处理，信噪比提高了12 dB以上。研究结果为激光声磁检测技术在钢轨无损检测工程领域的应用提供了一定的参考。

The objective of the study is to provide a novel method and system that is used to inspect the fatigue defects of the rail in a non-contact way. A laser-electro magnetic acoustic transducer (EMAT) system is designed in Laser Institute of Shandong Academy of Sciences, which is used in the present work. Artificial surface cracks and holes with different sizes are prepared on the rail. The principle of mode conversion is applied to determine the defects above. Namely, Rayleigh wave generated by line laser source becomes shear wave when the Rayleigh wave passes the surface defect. To receive the defect signal in a noncontact way, an EMAT sensor is used in this work. Experimental results show that the amplitude of defect signal increases with the increase of defect depth. To avoid the influence of noise on the testing result, the ultrasonic signal of surface defect is processed by wavelet soft-threshold filter designed by Matlab. The signal to noise ratio increases by 12 dB at least. The results show that the laser-EMAT technique has good prospects in the engineering applications of rail testing.