Location

Brunswick, ME

Start Date

1-1-1992 12:00 AM

Description

Order statistic and morphological filters belong to a class of nonlinear filters that have recently found many applications in signal analysis and image processing. In this paper, order statistic and morphological filters have been applied to enhance the features of the ultrasonic signal when it has been contaminated by multiple interfering microstructure echoes with random amplitudes and phases. These interfering echoes (i.e., speckles or grain scattering noise) often become significant to the point where detection of flaw echoes becomes very difficult. We have examined frequency diverse order statistic and time domain morphological and recursive median filters for improved ultrasonic flaw detection. In particular, the performance of these filters is evaluated using different ranks of order statistics and different shapes of structuring elements in the application of morphological filters. The processed experimental results in testing steel samples demonstrate that these filters are capable of improving flaw detection in ultrasonic systems.

Book Title

Review of Progress in Quantitative Nondestructive Evaluation

Volume

11A

Chapter

Chapter 3: Interpretive Signal Processing and Image Reconstruction

Section

Signal Processing

Pages

983-990

DOI

10.1007/978-1-4615-3344-3_126

Language

en

File Format

application/pdf

Share

COinS
 
Jan 1st, 12:00 AM

Ultrasonic Signal Enhancement Using Order Statistic and Morphological Filters

Brunswick, ME

Order statistic and morphological filters belong to a class of nonlinear filters that have recently found many applications in signal analysis and image processing. In this paper, order statistic and morphological filters have been applied to enhance the features of the ultrasonic signal when it has been contaminated by multiple interfering microstructure echoes with random amplitudes and phases. These interfering echoes (i.e., speckles or grain scattering noise) often become significant to the point where detection of flaw echoes becomes very difficult. We have examined frequency diverse order statistic and time domain morphological and recursive median filters for improved ultrasonic flaw detection. In particular, the performance of these filters is evaluated using different ranks of order statistics and different shapes of structuring elements in the application of morphological filters. The processed experimental results in testing steel samples demonstrate that these filters are capable of improving flaw detection in ultrasonic systems.