中文摘要：

在超声检测中,往往需要获得传播时间（TOF）、回波个数、中心频率、幅值等全面的信息,来综合评判缺陷的位置、大小和类型。通过建立多回波的卷积模型和参数化模型,给出一种结合最小熵盲反卷积（MED）和期望值最大（EM）算法思想的超声回波参数估计方法。首先基于卷积模型,采用最小熵反卷积,实现了重叠多回波信号的有效分离;再基于参数化模型和所获取的回波个数,给出了基于期望值最大算法思想的参数估计算法;最终实现了重叠多回波超声信号TOF、回波个数、中心频率、幅值等参数的精确估计。仿真和实验验证了该方法的有效性和优点。

英文摘要：

In ultrasonic testing,the parameters such as time of flight,number of echo,center frequency and amplitude are often necessary and primary for the accurate evaluation of type,size and location of defects. By establishing convolution model and parameterized model of multi-echoes,a parameters estimation method of ultrasonic echo is proposed,which combines minimum entropy deconvolution with the EM algorithm. Based on the convolution model,the separation of overlapped signals is realized by minimum entropy deconvolution. With the parameterized model and the number of echo,a parameters estimation method of he ultrasonic echo is proposed,and the parameters of the signal are precisely estimated. The effectiveness and advantage of the proposed method are verified through simulation signal and experiment.