New automatic image registration based on multi-features
YANG Meng , ZHU Chang-ren
(National Key Laboratory of Automatic Target Recognition, College of Electronic Science & Engineering, National University of Defense Technology, Changsha410073,China)
Abstract: This paper presented a new automatic image registration based on multi-features. The characteristic of this approach was that the invariant moment shape descriptor was used to establish correspondences between the potentially matched regions detected from the two images, and the improved signature was used to find the registration control points(RCP) as the stage of coarse registration. Then the improved chain-code matching had been performed as the registration refinement. Finally, the registration parameters were computed by using the least mean square. Experimental results with various kinds of image data show the high accuracy of our algorithm in multi-sensor image registration. ......