摘要
魏湘曙博士通过引进对偶空间,发现了二维图形制约三维形态的四个基本约束条件。利用它们构造相应的求解算子可以对二维线画图进行递推求解,但是,所求得的解对数值计算误差和结点位置偏差非常敏感,常常会导致线画图对应的多面体不存在。因此,误差校正是一个极其重要的问题,我们通过引进一个新约束条件设计了一个误差校正算法。对结点位置有偏差的线画图也能恢复多面体三维形态,并用实验验证了该算法的有效性。
By introducing a three-dimensional dual space, four basic constraints between polyhedron and its central projection image have been found by Dr. Wei. Using them, we can recover three -dimentional shapes of polyhedrons from line drawimg. But the solutions are sensitive to the errors in numerical calculation and vertex position errors, which will result in none possible polyhedron corresponding to a given line drawing. So. it is important to correct the errors. In this paper, we develop a practical method to correct the errors. A new constraint is given and an algorithm is designed. It can recover the polyhedron even if image data are incorrect due to vertex-position errors. An experiment is given which demonstrates the algorithm to be effective and efficient.
出处
《模式识别与人工智能》
EI
CSCD
北大核心
1994年第3期227-233,共7页
Pattern Recognition and Artificial Intelligence
关键词
机器视觉
线画图
三维
图像恢复
Computer Vision, Line Drawing, Dual Space, Correction
