Survey: State of the Art in NDE Data Fusion Techniques

TLDR

Nondestructive testing demands higher accuracy, prompting the use of multiple NDE methods and data fusion to combine multi‑source information and reduce uncertainty. The paper reviews progress in NDE data fusion techniques and examines various mathematical fusion algorithms. The authors describe a generic framework for applying NDE data fusion, reviewing optimization, multiresolution, heuristic, probabilistic, and visualization algorithms. Fusion of advanced NDE and sensor techniques promises new diagnostic and prognostic health‑monitoring applications.

Abstract

As the requirements on the accuracy of nondestructive testing and evaluation are increasing, multiple nondestructive evaluation (NDE) methods are often employed to increase the reliability and reduce the uncertainty of the testing and evaluation. The need to process and analyze data from multiple sources uses the technique named data fusion, which has been growing rapidly in the NDE community in recent years. This paper reviews the progress in NDE data fusion techniques and examines the mathematical fusion algorithms, which include but are not limited to optimization methods, multiresolution approaches, heuristic methods, probabilistic methods, and scientific visualization. In the light of recent applications in NDE data fusion, important issues in handling the acquired data and applying the fusion algorithms are identified and discussed. A generic framework to apply the NDE data fusion process is described. The study on fusion of advanced NDE and sensor techniques opens up prospects for the diagnostic and prognostic health monitoring applications in the future.