Apply Neutron Radiation and X-Ray Images Fusion for Defect Detection in Manufacturing

Electronic products defect detection is an important branch of electronic image processing, in which, image fusion can be applied to deal with the task. In view of contourlet is more suitable to describe anisotropic lines, curves and boundary characteristics of image than wavelet, and can effectively capture the geometry structure. Comparing with contourlet transform, non-subsampled contourlet transform (NSCT) not only enjoys features of multi-scale, multi-direction, but also possesses translational invariance. In this paper, a new digital neutron and X-ray images fusion method based on contourlet and IHS transformation is proposed. The results show that our method preserves more image detail and hence provides more accurate information, which is superior to IHS and wavelet-IHS based fusion approaches. It is proved to be an effective approach to industrial production.