Description

Abstract:

          In this paper, three image registration methods based on feature matching has been

described and the correlation experiment contrast has been processed. Harris method is good at its

simple calculation and evenly reasonable angular point feature extraction, but is ineffective at

matching effect of image rotation scale variation.  SIFT method is slowest, but it has the

advantaged of affine, rotating, scale invariable. The advantage of SURF method lies in its fastest,

and is able to Realize real-time requirements, at the same time its matching effect is basically

equal to SIFT method, but it is bad to color and illumination change. Feature matching algorithm:

Harris corner detection algorithm, SIFT (Scale Invariant Feature Transform) algorithm, SURF

（Speeded-Up Robust Features ） algorithm.