Separation Study of Ultrasonic Mixed Signals Based on Linear Radon Transform

In the process of thickness measurement of large-thickness media using one-shot low-frequency reflection ultrasound method, the collected reflection ultrasound signal has the phenomenon of effective wave and direct wave mixing, which is difficult to be processed effectively by conventional one-dimensional ultrasound signal processing methods. In order to solve this kind of signal mixed problem, an adaptive phase subtraction processing method based on linear Radon transform was proposed. From the two-dimensional perspective, according to the characteristics that direct interference shows regular consistency and effective underside echo shows shift difference in time domain, the signal array was converted to linear Radon domain. In order to improve the resolution of low-frequency ultrasonic echo data, adaptive phase subtraction algorithm was applied to the signal data after projection convergence to effectively suppress and separate the direct interference waves. Finally, the effectiveness of the method is verified by simulated data processing, and the relative error of the corresponding thickness solution does not exceed 0.7%. According to the simulation and practical application results, the proposed method can well solve the separation problem of this kind of mixed ultrasonic signals.