Novel S-transform information fusion for filtering ultrasonic pulse-echo signals

Direct evaluation of ultrasonic signals requires data analyses with an acceptable level of noise. Ultrasonic signals represent a specific category of time domain signals to be analyzed. In order to increase a difference between the level of noise and the amplitude of the ultrasonic pulse a suitable method for signal filtering has to be used. Within this article we discuss and evaluate a novel signal denoising method. The S-transform for signal analysis and processing was used. This transformation has been recently introduced for ultrasonic echo analyses. Proposed transformation represents an intermediate stage between the Fourier transform analysis and the wavelet transform analysis. In order to filter ultrasonic signals from the Electromagnetic Acoustic Transducer (EMAT) with a high level of noise, new, different approach in signal filtering was developed based on an information fusion. Suggested method is able to process the pulse-echo signal in its full complexity. Proposed method offers good results in studied ultrasonic signals in comparison to digital filter or wavelet denoising.