Angle Correction Algorithm for Image Rotation Based on Key Point Detection

In image matching, traditional template matching method is not suitable for matching two images with large rotation angle. In order to solve this promble, this paper proposes an angle correction algorithm based on key point detection,which begins with corner detection by corner detector based on adaptively filtered curvature function. Then, a stable key piont is selected and the angle of the rotation image is computed between the selectded point and the geometric center of the image. Contrasting to the existing algorithms that solve rotation angle, the proposed algorithm can solve the promble of arbitrary angle rotation and has a high calibration accuracy and applicability. Experimental results demonstrate the availability and advantages of the algorithm.