Ultrasonic surface parameterisation using maximimum entropy signal processing

P. F. Smith, M. A. Player

Research output: Contribution to journalArticlepeer-review

Abstract

This paper demonstrates the value of the Maximum Entropy Method (MEM) for assisting in the estimation of surface amplitude distributions from reflected ultrasonic pulses. Roughnesses ranging from around 2μm to over 140μm (R) have been measured using LMN transducers with centre frequencies ranging from 10 to 40 MHz. Recent enhancements to the maximum entropy method give improved results, with higher resolution and a capability of quantifying an error for each deconvolution image. A selection of results from rough and layered surfaces is presented comparing the new MEMSYS 3 results with those from the original MEM and those measured by the Form Talysurf.

Original languageEnglish
Pages (from-to)25-42
Number of pages18
JournalNondestructive Testing and Evaluation
Volume10
Issue number1
DOIs
Publication statusPublished - 1 Jul 1992

Keywords

  • Surface parameterisation
  • Ultrasonic signal processing
  • Ultrasonic transducers

Fingerprint

Dive into the research topics of 'Ultrasonic surface parameterisation using maximimum entropy signal processing'. Together they form a unique fingerprint.

Cite this