A statistical image fusion algorithm for NDE applications

Data fusion techniques have been used extensively in several application areas, including nondestructive evaluation (NDE) applications. This paper proposes a probability density based pixel-level fusion algorithm for NDE applications. The proposed approach models the fusion process as a convex combination of input data. The model parameters are estimated by optimizing the Kullback-Leibler distance between the density functions of the fused and the input data. Preliminary results on multimodal eddy current NDE data from aircraft lap joints indicate the feasibility of the proposed approach.