• DocumentCode
    2315622
  • Title

    A Hybrid Algorithm for Solving 7 Parameters Transformation

  • Author

    Hashemi, Amir ; Kalantari, Mahzad

  • Author_Institution
    Dept. of Math. Sci., Isfahan Univ. of Technol., Isfahan, Iran
  • fYear
    2009
  • fDate
    26-29 Sept. 2009
  • Firstpage
    37
  • Lastpage
    44
  • Abstract
    The 7 parameters transformation is a well-known problem in engineering sciences such as Computer Vision, Survey Engineering and Photogrammetry. To solve this problem, we use an algebraic solver in which we need to transform the input system into an equivalent system, but better adapted such as a Gro¿bner basis. For operational applications of this problem, we need to compute this basis as fast as possible. In this paper, we describe an efficient hybrid algorithm which employs both Buchberger algorithm and Faugere\´s F5 algorithm to compute this basis. Using this algorithm, we record the "trace" of the useful calculations for computing the Gro¿bner basis of the polynomial ideal generated by the equations of the system. For any coordinates of input points, this trace can provide directly the Gro¿bner basis of the ideal generated by the system without useless computations.
  • Keywords
    nonlinear equations; polynomials; process algebra; 7 parameters transformation; Buchberger algorithm; Faugere F5 algorithm; Gro¿bner basis; algebraic solver; computer vision; engineering science; equivalent system; hybrid algorithm; photogrammetry; survey engineering; Application software; Computer vision; Equations; Geodesy; History; Jacobian matrices; Polynomials; Scientific computing; Stereo vision; Transforms; 7 parameters transformation problem; Buchberger algorithm; Faugère´s F5 algorithm; Gröbner basis; computer vision; direct solution; photogrammetry;
  • fLanguage
    English
  • Publisher
    ieee
  • Conference_Titel
    Symbolic and Numeric Algorithms for Scientific Computing (SYNASC), 2009 11th International Symposium on
  • Conference_Location
    Timisoara
  • Print_ISBN
    978-1-4244-5910-0
  • Electronic_ISBN
    978-1-4244-5911-7
  • Type

    conf

  • DOI
    10.1109/SYNASC.2009.17
  • Filename
    5460869