The Study Of Image Stitching Algorithm Based On The Feature

One important step in the technique of image processing is Image Stitching. It can be divided into two classes: the area method and the feature method. The normalized cross correlation algorithm is in common use method of the area matching. Since in the case of dissimilar image matching, the matching probability of correlation algorithm is not good. Recent many studies are focus on feature matching.When images have significant differences in translation, rotation or scale and blurred by noises, most of multimodal registration methods is not satisfied by right matching. Feature-based methods perform well for the images having large misalignment, and their computational costs are low.Feature matching includes two steps. One is the select of feature and feature detection, another is the matching measure. To rotation images, which are blurred by noise, SUSAN and Harris are compared in detect precision and compute efficiency. SUSAN is used for its noise and rotation insensitivity and little time consuming.Based on the corner detection algorithm, an image automatic registration method, which to rotation and blurred by noise, was developed. In the corner matching procedure, the correspondence between the corner points in two images was established with three steps.First, a primary correspondence was initialized using normalized cross-correlation. Secondly, the wrong match points were deleted based on the Mahalanobis distance with affine invariance. Thirdly, the three corner couples were select by normalized correlation. Finally, the affine transformation parameters were estimated from the correctly corresponding points.Our Stitching method works well for the images with significant translation and rotation. It is suitable for automatic registration between the images with less noisy and scale variance. Many experiments are implemented both on corner detector and feature matching in this paper. Experiments demonstrated the effectiveness of the method in both accuracy and computational speed.