X-ray image mosaic based on scale invariant feature extraction

Image mosaic technology based on scale invariant is proposed in this paper for industrial X-ray image with low contrast and some artefacts. Firstly, scale-invariant feature point extraction algorithm uses of to extract the feature points. Then, the coarse feature points matching based on maximum mutual information, and fine matching based on RANSAC algorithm to removing the fake matching pairs. Experimental results show that the algorithm has good stability, especially with 20 percent Gaussian noise, this algorithm can still be detected feature points actually. Compared with other feature point matching methods, the algorithm maintain stability at the same time improve the matching accuracy.