An improved registration method using the criterion of five-coplanar points

An improved registration method using the criterion of five-coplanar points has been proposed. Registration precision depends on the accuracy of feature matching pairs, so how to refine the matching pairs is very important in the process of image registration. The scale invariant feature transform (SIFT) detector have been used to extract the feature matching pairs and the random sample consensus (RANSAC) is used to remove the outliers. The measure based on five coplanar points structure has been employed refine the feature matching pairs. Our experiment shows that the alignment error of our method is much smaller than that of normal method. The ascension of registration precision is obvious. The registration accuracy and stability of our method are still better than that of normal method under the degradation of blurring and noise.