Title :
Automatic reconstruction of 3D branching objects
Author :
Sirakov, Nikolai Metodiev
Author_Institution :
Inst. of Mech. & Biomech., Bulgarian Acad. of Sci., Sofia, Bulgaria
Abstract :
In this paper we assume that 3D objects are known as a database containing 3D co-ordinates and gray tone level of the object points. With the help of the database, we construct a sequence of planes containing a set of 2D objects. In order to reconstruct several 3D branching objects we solve the following problems: 1) divide the set of 2D objects to subsets of similar 2D objects; 2) find the branches of similar 2D objects; and 3) connect similar 2D objects and branches. We introduce the term morphological similarity to solve the first problem. We apply morphological sequences to find the branches. We determine the “corresponding” essential points to connect similar 2D objects. We performed experiments to validate the theory, using real data obtained from ore deposit
Keywords :
computer vision; image reconstruction; mathematical morphology; mineral processing industry; mining; stereo image processing; visual databases; 3D branching objects; computer vision; database; gray tone level images; image reconstruction; mining industry; morphological sequences; morphological similarity; Biomechanics; Humans; Minerals; Mining industry; Neoplasms; Object detection; Ores; Production; Prosthetics; Shape;
Conference_Titel :
Pattern Recognition, 1996., Proceedings of the 13th International Conference on
Conference_Location :
Vienna
Print_ISBN :
0-8186-7282-X
DOI :
10.1109/ICPR.1996.546898