Abstract

Our research addresses manual inspection and automatic registration of NDE data in the context of the digital thread / digital twin via inspection location tracking. Here, manual inspection refers to all tasks where an inspector acquires NDE data manually using a handheld probe. Automatic data integration is challenging in this situation since this task requires spatial position and orientation (pose) data for each NDE measurement. We investigate 3D machine vision methods for pose estimation. In brief, an RGB-D camera observes the asset under inspection along with the probe; 3D machine vision processes the camera data to actively track the probe in relation to the asset, which further allows one to augment each NDE dataset with its inspection location. This location facilitates to automatically integrate the data into a shape model of a digital twin. We already presented a system prototype focusing on flash thermography, which successfully addressed challenges affecting the accuracy and robustness.

How to Cite:

Radkowski, R., Garrett, T. & Holland, S. D., (2019) “3D machine vision technology for automatic data integration of ultrasonic data”, Review of Progress in Quantitative Nondestructive Evaluation .