作者: David Cyganski , T.Andrew Cott , John A. Orr , Reynold J. Dodson
DOI: 10.1117/12.942717
关键词: Affine transformation 、 Affine invariant 、 Mathematics 、 Planarity testing 、 Computer vision 、 Image matching 、 Robustness (computer science) 、 Invariant (mathematics) 、 Artificial intelligence 、 Affine shape adaptation 、 Curvature
摘要: New methods of image matching and calculation the affine transformation relating two images an object in different orientations are developed. These applicable to contours extracted from planar-patch objects. The approach is based on a new definition curvature function (previously developed by authors) which invariant all distortions representable transformations. algorithm implemented tested using camera-acquired actual objects, seen exhibit considerable robustness real-world imaging process, as well deviations planarity