Detection of exact and near duplicates in phased-array ultrasound weld scan

Abstract

Finding duplicates in phased-array ultrasound weld scans is typically done by certified analysts. Due to the vast amount of data involved, this task is time-consuming and error-prone. Some automated analysis software exists but are still inefficient. In this paper, two types of duplicates are defined. A method to detect each type in a database of three-dimensional ultrasound weld scans is presented. The duplicate detection process is implemented in a cloud-computing service enabling near real-time processing in industrial applications. Experiments are performed to expose the detection accuracy of both methods.