Three-dimensional Reconstruction of Rotated Object Based on Contour Lines
Subin Qian · Journal of Chengdu University · 2010
Three-dimensional reconstruction from two-dimensional grey image of rotated object has been a research hot spot in the filed of computer vision.Based on the deep research on it,a method was proposed which applied message feature of contour lines from single image to reconstruct rotated object.The method extracted the rotated object as the research target,adopted corner detection to acquire valid contour lines and their sub-curves.Then it acquired the axis of rotated object according to the type of sub-curves and information of characteristic points and combined the geometric properties of rotated object to realize the three-dimensional reconstruction of curved surface.The experiment indicates that the method can conveniently reconstruct the three-dimensional shape of the rotated object according to its two-dimensional grey image.