On the Effect of Control-Point Spacing on the Multisolution Phenomenon in the P3P Problem

The perspective 3-point (P3P) problem, also known as pose estimation, has its origins in camera calibration and is of importance in many fields: for example, computer animation, automation, image analysis, and robotics. One possibility is to formulate it mathematically in terms of finding the solution to a quartic equation. However, there is yet no quantitative knowledge as to how control-point spacing affects the solution structure—in particular, the multisolution phenomenon. Here, we consider this problem through an algebraic analysis of the quartic’s coefficients and its discriminant and find that there are significant variations in the likelihood of two or four solutions, depending on how the spacing is chosen. The analysis indicates that although it is never possible to remove the occurrence of the four-solution case completely, it could be possible to choose spacings that would maximize the occurrence of two real solutions. Moreover, control-point spacing is found to impact significantly on the reality conditions for the solution of the quartic equation.

[1]  A. D. N. Smith.  THE EXPLICIT SOLUTION OF THE SINGLE PICTURE RESECTION PROBLEM, WITH A LEAST SQUARES ADJUSTMENT TO REDUNDANT CONTROL , 2006 .

[2]  Michael Q. Rieck.  A Fundamentally New View of the Perspective Three-Point Pose Problem , 2013, Journal of Mathematical Imaging and Vision.

[3]  Michael Q. Rieck.  Handling Repeated Solutions to the Perspective Three-point Pose Problem , 2010, VISAPP.

[4]  Michael Vynnycky,et al.  Mathematical Analysis of the Multisolution Phenomenon in the P3P Problem , 2014, Journal of Mathematical Imaging and Vision.

[5]  胡占义,et al.  A General Sufficient Condition of Four Positive Solutions of the P3P Problem , 2005 .

[6]  G. Volkart.  Das Rückwärtseinschneiden im Raum , 1933 .

[7]  Michael J. Magee,et al.  The Perspective View of Three Points , 1991, IEEE Trans. Pattern Anal. Mach. Intell..

[8]  Robert C. Bolles,et al.  Random sample consensus: a paradigm for model fitting with applications to image analysis and automated cartography , 1981, CACM.

[9]  Hu Zhan,et al.  A Probabilistic Study on the Multiple Solutions of the P3P Problem , 2007 .

[10]  E. L. Rees,et al.  Graphical Discussion of the Roots of a Quartic Equation , 1922 .

[11]  Su Chen.  NECESSARY AND SUFFICIENT CONDITION OF POSITIVE ROOT NUMBER OF PERSPECTIVE-THREE-POINT PROBLEM , 1998 .

[12]  Jifu Zhang,et al.  Singular Points in the Optical Center Distribution of P3P Solutions , 2015 .

[13]  Robert M. Haralick,et al.  Review and analysis of solutions of the three point perspective pose estimation problem , 1994, International Journal of Computer Vision.

[14]  Michael Q. Rieck.  Related Solutions to the Perspective Three-Point Pose Problem , 2015, Journal of Mathematical Imaging and Vision.

[15]  Jianliang Tang,et al.  Complete Solution Classification for the Perspective-Three-Point Problem , 2003, IEEE Trans. Pattern Anal. Mach. Intell..

[16]  Joseph S.-C. Yuan.  A general photogrammetric method for determining object position and orientation , 1989, IEEE Trans. Robotics Autom..

[17]  Larry S. Davis,et al.  Pose Determination of a Three-Dimensional Object Using Triangle Pairs , 1988, IEEE Trans. Pattern Anal. Mach. Intell..

[18]  Hu,et al.  Why is the Danger Cylinder Dangerous in the P3P Problem , 2006 .

[19]  Roland Siegwart,et al.  A novel parametrization of the perspective-three-point problem for a direct computation of absolute camera position and orientation , 2011, CVPR 2011.

[20]  Michael Q. Rieck.  An Algorithm for Finding Repeated Solutions to the General Perspective Three-Point Pose Problem , 2011, Journal of Mathematical Imaging and Vision.

[21]  Daniel Lazard,et al.  Quantifier Elimination: Optimal Solution for Two Classical Examples , 1988, J. Symb. Comput..

[22]  Lu Yang.  A Simplified Algorithm for Solution Classification of the Perspective-three-Point Problem 1) , 1998 .