Title :
Improved OBB fitting method for elongated objects
Author :
Ma, Gang ; Jia, Jinyuan ; Zhang, Qian
Author_Institution :
Sch. of Software Eng., Tongji Univ., Shanghai, China
Abstract :
We present a new method of calculating OBBs of the internal nodes in OBB-tree for elongated objects which are partially ordered. By establishing a new estimating function, we convert the problem of determining OBB´s orientation to constrained extrema problem which can be solved by Lagrange multiplier method. After solving the estimating function we get a local axis of the OBB. The remaining two directions are calculated using minimal-area bounding rectangle algorithm. Project all the vertices to the local axes respectively then the size and center of the OBB can be finally determined. A large amount of tests have revealed that the OBB-tree constructed by this method fits better than those constructed by the already existing methods. Now the method proposed by this paper has been adopted by actual project.
Keywords :
computational geometry; curve fitting; trees (mathematics); Lagrange multiplier method; OBB fitting method; OBB-tree; elongated objects; minimal-area bounding rectangle algorithm; Costs; Covariance matrix; Engine cylinders; Information geometry; Iterative algorithms; Lagrangian functions; Road transportation; Software engineering; Solid modeling; Testing; Lagrange multiplier; OBB; major direction; minimal-area bounding rectangle;
Conference_Titel :
Computer-Aided Industrial Design & Conceptual Design, 2009. CAID & CD 2009. IEEE 10th International Conference on
Conference_Location :
Wenzhou
Print_ISBN :
978-1-4244-5266-8
Electronic_ISBN :
978-1-4244-5268-2
DOI :
10.1109/CAIDCD.2009.5375284