Data Fusion of Eddy Current, Ultrasonic, and Radiographic Data

The availability of large area robotic scanners for nondestructive inspection (NDI) techniques such as ultrasound (UT) or eddy current (ET) and intrinsically area-based techniques such as digital radiography (DR) or thermography have changed the inspection process in the depot and field. Texas Research Institute Austin, Inc. (TRI/Austin), in collaboration with Computational Tools Inc., Victor Technologies LLC. (VTL), and The Boeing Company, propose to develop a software tool and associated database structure to perform application-optimized data fusion and mining for NDI data. The TRI team will 1) develop the capability to perform automated defect detection and classification using individual and multiple (i.e. before and after repair, depot cycle interval) automated UT or DR inspections of the B-1B horizontal stabilizer, using existing data to be supplied by the AF, 2) develop the requirements for databases to enable the implementation of these data analysis/fusion algorithms along with data mining in the high throughput environment of the AF depot, 3) document how the basic analysis/fusion capability developed in the first goal can be extended to additional applications of interest to the USAF, such as general corrosion-induced material thinning in multilayered aircraft structures.