Wavelet based deconvolution method in ultrasonic tomography

Philippe Lasaygues (1)

1 LMA - Laboratoire de Mécanique et d'Acoustique [Marseille]

This paper deals with the quantitative and qualitative ultrasonic imaging of media contrasting strongly with the background (bones, for example). The wave propagation in these structures, which involves on several phenomena, generates ultrasound signals forming packages with different time and frequency signatures. To correlate the complex physical processes involved with the various packets forming the signals, it is advisable to separate the useful information (about the various packages, for example) from the parasite information (electronic noise). After testing several possible methods (filtering, deconvolution) with variable levels of success, it was proposed here to increase the resolution of the ultrasound signals, using a method of processing which was optimized by simultaneously analyzing the time and frequency parameters. For this purpose, algorithms based on the wavelet decomposition of the signals were used and suitable transmitted signals correlated with the parameters of the experimental device and the incident wave, were selected. The mathematical properties of the transmitted signal could be adapted themselves well to this time and frequency based approach. This method, which we have called the "wavelet based deconvolution method", makes it possible to determine the transfer function of the scatterer, which we attempted to reconstruct.

Domaines: Acoustique [physics.class-ph]; Acoustique [physics.class-ph]

Philippe Lasaygues. Wavelet based deconvolution method in ultrasonic tomography. International congress on ultrasonics, 2007, Vienne, Austria. CD-ROM(4 p.). ⟨hal-00440745⟩