Robust Homography Estimation from Local Affine Maps

The corresponding point coordinates determined by classic image matching approaches define local zero-order approximations of the global mapping between two images. But patches around keypoints typically contain more information, which may be exploited to obtain a first-order approximation mapping, incorporating affine maps keypoints. Several methods have been proposed in literature compute this approximation. In paper we present several modifications RANSAC (RANdom SAmple Consensus) algorithm, that uses and a-contrario procedures improve homography estimation pair methodology provides definition soundness an allows for adaptive thresholds inlier/outlier discrimination. These outperform state-of-the-art different choices descriptors datasets, permit increase probability success identifying pairs challenging databases.