Fusion of multi-view ultrasonic data for increased detection performance in non-destructive evaluation

Area of Science:

• Materials Science
• Engineering
• Non-Destructive Evaluation

Background:

• State-of-the-art ultrasonic non-destructive evaluation (NDE) employs arrays for rapid, multi-view data acquisition on critical components.
• Interpreting dispersed defect information across numerous views necessitates extensive operator analysis.

Purpose of the Study:

• To develop and evaluate methods for fusing multi-view ultrasonic NDE data.
• To enhance defect detection performance and reduce operator burden in NDE.

Main Methods:

• Development of various multi-view data fusion techniques for ultrasonic NDE.
• Comparative analysis of fusion methods against single-view performance.

Main Results:

• All fusion methods demonstrated significant performance gains over individual views.
• Fusion effectively utilizes information from the most suitable views, reducing operator workload.
• Specific fusion techniques improved detection of challenging defects with fewer false alarms.

Conclusions:

• Multi-view data fusion substantially enhances ultrasonic NDE performance for both general and edge-case defect detection.
• Fusion methods offer a pathway to extend NDE capabilities, impacting engineering asset design and operation.