Automated segmentation of ultrasonic volumetric data of composite materials

Ahmad Osman (1) , Ulf Hassler (2) , Valerie Kaftandjian (3) , Joachim Hornegger (4)

1 Fraunhofer IZFP - Fraunhofer Institute for Nondestructive Testing
2 Fraunhofer IIS/EZRT - Fraunhofer-Entwicklungszentrum Röntgentechnik
3 LVA - Laboratoire Vibrations Acoustique
4 FAU - Friedrich-Alexander Universität Erlangen-Nürnberg = University of Erlangen-Nuremberg

Manual ultrasonic testing is frequently applied to inspect the quality of carbon fibre-reinforced polymer composites (CFRP), especially in the aerospace industry. There is an immediate requirement to replace manual ultrasonic testing and data evaluation with an automated system. A part of the solution is provided via automated acquisition of ultrasound data. The second part to be solved is the analysis of the enormous amount of data generated. The objective of this contribution is to propose a method that can provide an automated interpretation of data. The proposed method is tested on two CFRP samples with artificial defects of different sizes, depths and shapes.

Mots clés: sampling phased array; composite materials; ultrasound image analysis; speckle noise reduction

Domaines: Sciences de l'ingénieur [physics]

Ahmad Osman, Ulf Hassler, Valerie Kaftandjian, Joachim Hornegger. Automated segmentation of ultrasonic volumetric data of composite materials. Insight - Non-Destructive Testing & Condition Monitoring, 2015, 57 (3), ⟨10.1784/insi.2014.57.3.153⟩. ⟨hal-01282089⟩