Method of dynamic thresholding for flaw detection in ultrasonic C-scan images
US5608814A · kind A · utility
Assignee
Inventors
Key dates
| Filing date | Nov 17, 1994 |
| Grant date | Mar 4, 1997 |
| Priority date | — |
| Expiry date | Nov 17, 2014 |
Classification
- Technology area (CPC G)Physics
- CPC primaryG06V10/28
- WIPO fieldComputer technology
- WIPO sectorElectrical engineering
Abstract
The method of producing a binary flaw-no flaw image of an object, including employing an ultrasonic data acquisition system to obtain data values f(i,j) which define a C-scan image (F) of the object, dividing the C-scan image (F) into a plurality of subimages (G.sub.k for k=1,2, . . . ,K), determining regional threshold levels y(k) for each of the plurality of subimages, using said regional threshold levels y(k) to determine pixel threshold levels t(i,j) for each pixel (i,j) of the image (F) by interpolation, and generating a binary flaw-no flaw image (B) by assigning binary values thereto based on a comparison between the pixel threshold levels t(i,j) and data values f(i,j), thereby providing a method which achieves a high probability of flaw detection and a low probability of false flaw indications.
Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.