Heuristic Understanding of Three Orthographic Views

Access this Article

Search this Article

Author(s)

    • Changhun Kim
    • Institute of Information Sciences and Electronics, University of Tsukuba
    • Masahiro Inoue
    • Institute of Information Sciences and Electronics, University of Tsukuba

Abstract

This paper describes a system that reconstructs 3-D models of polyhedrons from three orthographic views represented by binary images. Our main stress is on the description of a new reconstruction algorithm. The pro-posed algorithm performs a combinatorial search based on the face decision strategy along with two heuristics. The constraint rules are developed on the basis of the properties of orthogonal projection geometry which are dynamically applied to the combinatorial search process. Further two heuristics are adopted in the combinatorial search process so that its way of interpretation follows that of human understanding. One is for ordering the search nodes to reduce the search space and the other is for reconstructing the scenes as naturally as possible. The effectiveness of introducing heuristics is also proved and analyzed experimentally. It is shown that 3-D scenes can successfully be reconstructed by the proposed algorithms and that the efficiency of the understanding system is improved considerably by the introduction of the heuristics.This paper describes a system that reconstructs 3-D models of polyhedrons from three orthographic views represented by binary images. Our main stress is on the description of a new reconstruction algorithm. The pro-posed algorithm performs a combinatorial search based on the face decision strategy along with two heuristics. The constraint rules are developed on the basis of the properties of orthogonal projection geometry, which are dynamically applied to the combinatorial search process. Further, two heuristics are adopted in the combinatorial search process so that its way of interpretation follows that of human understanding. One is for ordering the search nodes to reduce the search space, and the other is for reconstructing the scenes as naturally as possible. The effectiveness of introducing heuristics is also proved and analyzed experimentally. It is shown that 3-D scenes can successfully be reconstructed by the proposed algorithms, and that the efficiency of the understanding system is improved considerably by the introduction of the heuristics.

This paper describes a system that reconstructs 3-D models of polyhedrons from three orthographic views represented by binary images. Our main stress is on the description of a new reconstruction algorithm. The pro-posed algorithm performs a combinatorial search based on the face decision strategy along with two heuristics. The constraint rules are developed on the basis of the properties of orthogonal projection geometry, which are dynamically applied to the combinatorial search process. Further, two heuristics are adopted in the combinatorial search process so that its way of interpretation follows that of human understanding. One is for ordering the search nodes to reduce the search space, and the other is for reconstructing the scenes as naturally as possible. The effectiveness of introducing heuristics is also proved and analyzed experimentally. It is shown that 3-D scenes can successfully be reconstructed by the proposed algorithms, and that the efficiency of the understanding system is improved considerably by the introduction of the heuristics.

Journal

  • Journal of Information Processing

    Journal of Information Processing 15(4), 510-518, 1992-12-31

    Information Processing Society of Japan (IPSJ)

Codes

  • NII Article ID (NAID)
    110002673669
  • NII NACSIS-CAT ID (NCID)
    AA00700121
  • Text Lang
    ENG
  • Article Type
    Article
  • ISSN
    1882-6652
  • Data Source
    NII-ELS  IPSJ 
Page Top