This research addresses inspection location tracking in the field of nondestructive evaluation (NDE) using a computer vision technique to determine the position and orientation of typical NDE equipment in a test setup. The objective is the tracking accuracy for typical NDE equipment to facilitate automatic NDE data integration. Since the employed tracking technique relies on surface curvatures of an object of interest, the accuracy can be only experimentally determined. We work with flash-thermography and conducted an experiment in which we tracked a specimen and a thermography flash hood, measured the spatial relation between both, and used the relation as input to map thermography data onto a 3D model of the specimen. The results indicate an appropriate accuracy, however, unveiled calibration challenges.
Available at: http://works.bepress.com/robert_grandin/18/
This proceeding may be downloaded for personal use only. Any other use requires prior permission of the author and AIP Publishing. This proceeding appeared in Radkowski, Rafael, Stephen Holland, and Robert Grandin. "Evaluation of the fidelity of feature descriptor-based specimen tracking for automatic NDE data integration." AIP Conference Proceeding 1949, no. 1 (2018): 030008, and may be found at DOI: 10.1063/1.5031531. Posted with permission.