• DocumentCode
    1952972
  • Title

    Deconvolution of ultrasonic signals in time domain

  • Author

    Gust, Norbert ; Kuehnicke, Elfgard

  • Author_Institution
    Tech. Univ. Dresden, Dresden, Germany
  • fYear
    2010
  • fDate
    11-14 Oct. 2010
  • Firstpage
    1817
  • Lastpage
    1820
  • Abstract
    Deconvolution can be used to separate overlapping echoes from very thin layers. This way the reflection parameters of each particular echo can be obtained. Here, a noise insensitive time-domain deconvolution algorithm is presented, which can handle arbitrary and time-of-arrival dependent reference pulse shapes. This can reduce artifacts and enable proper detection of echoes with small amplitude, even if echoes have a non-uniform shape like real echoes do. In contrast to common deconvolution approaches it does not assume Gaussian pulse shapes or the same reference shape for all reflections. The deconvolution results have a very high quality with artifacts in the lower 1% range. Pulses are separated at arrival time differences above half a wave-period. The high precision of the reflection parameter determination enables much better analysis quality also for applications with no superposition of reflections.
  • Keywords
    acoustic signal processing; acoustic wave reflection; deconvolution; echo; ultrasonics; arbitrary reference pulse shapes; artifact reduction; echo reflection parameters; noise insensitive time domain deconvolution algorithm; overlapping echo separation; reflection parameter determination; small amplitude echo detection; time domain ultrasonic signal deconvolution; time of arrival dependent reference pulse shapes; Acoustics; Deconvolution; Matching pursuit algorithms; Materials; Prediction algorithms; Reflection; Shape;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Ultrasonics Symposium (IUS), 2010 IEEE
  • Conference_Location
    San Diego, CA
  • ISSN
    1948-5719
  • Print_ISBN
    978-1-4577-0382-9
  • Type

    conf

  • DOI
    10.1109/ULTSYM.2010.5935519
  • Filename
    5935519