X-ray Image Sequences Based On Feature Matching Automatic Stitching Algorithm Research And Implementation

To measure the scoliosis and the lower limb deformity, to resolve the problem of the size of the body examined much more than that of the X-ray detector, and to provide data important to doctors, an automatic stitching method of an X-ray image-sequence based on biorthogonal wavelet transformation and feature matching has been proposed in this paper.First of all, the algorithm of each step was summarized and analysed, and the algorithms of the registration were studied and classified in this paper. Then, biorthogonal wavelet transformation was employed to implement decomposing of the multi-resolution, and the effective edge of the image can be extracted by this method combined with Canny operator. The feature points of the image can be obtained by calculating the edge contour matrix E and the value matrix H. The rough matching of feature points can be achieved by using normalized cross correlation(NCC) algorithm, and the random sample consensus (RANSAC) algorithm is introduced to remove false matching pairs and realize precise matching. The image-sequence is automatically ordered by the improved genetic algorithm to implement automatic stitching. The weighted average fusion algorithm is applied to achieve smooth and seamless stitching. Followingly, stitching experiments of several groups of X-ray image-sequence with gray discrepancy, contrast discrepancy, rotation discrepancy, and scale discrepancy were carried out to prove the stability of the method proposed in this paper. Finally, this method is applied in the clinical orthopedic to mosaic X-ray images. Experimental results show that this method can rapidly and accurately realize the X-ray image-sequence automatic stitching without external features, and therefore it can be applied to in medical imaging field widely.