Application of 2-dimensional analytic signals with single-quadrant spectra for processing of SAFT-reconstructed images

• Published in: Multidimensional systems and signal processing Vol. 25; no. 4; pp. 703 - 722
• Main Authors: Höhne, Christian, Boehm, Rainer, Prager, Jens
• Format: Journal Article
• Language: English
• Published: Boston Springer US 01.10.2014
• Subjects: Artificial Intelligence; Circuits and Systems; Electrical Engineering; Engineering; Signal,Image and Speech Processing; Complex signal; Image processing; SAFT image; Analytic signal
• ISSN: 0923-6082; 1573-0824
• DOI: 10.1007/s11045-013-0226-7

The Synthetic Aperture Focusing Technique (SAFT) is an algorithm applied in non-destructive ultrasonic testing which provides an image of flaws within a specimen. The image is reconstructed from A-scans measured at different positions. Reliable evaluation of the images obtained by the SAFT-algorithm, however, depends on the representation of the reconstructed data, which is initially given in terms of positive and negative local values only. A suitable way of processing this data for evaluation is to calculate the envelope, which can be achieved by means of the analytic signal. The extension of this concept to the multidimensional case is neither trivial nor unique and although extensive work on this subject has been carried out in the past, a correct envelope calculation in multidimensional data remains difficult since it depends on an additional condition, namely the separability of the signal. In this paper, the concept of analytic signals with single-quadrant spectra is applied to process 2-dimensional data obtained by the SAFT-algorithm. Furthermore, we present a procedure to overcome the limitations of that approach by selecting local magnitude values from a number of rotated frames after evaluating the signal’s separability in each frame, which is briefly validated against synthetic and experimental data.