Monocular vision pose estimation method based on coplanar target

• Main Authors: HU QINGLEI, ZHENG JIANYING, GUO LEI, JIANG CUICUI
• Format: Patent
• Language: Chinese; English
• Published: 10.09.2021
• Subjects: CALCULATING; COMPUTING; COUNTING; ELECTRIC DIGITAL DATA PROCESSING; IMAGE DATA PROCESSING OR GENERATION, IN GENERAL; PHYSICS

The invention discloses a monocular vision pose estimation method based on a coplanar target. The method comprises the following steps: defining a unit virtual spherical surface, and projecting coplanar feature points on pixel points of an image plane to the spherical surface; constructing the square of the distance between two points of the spherical surface as a target function, and obtaining local minimum values of two pose coupling terms by adopting a quadratic Levenberg-Marquardt nonlinear optimization method; and selecting a pose coupling item with a smaller objective function as global optimum, and obtaining a pose by using singular value decomposition. The method ensures that the global optimal pose can still be obtained under the condition that the coplanar feature point target function has multiple local minimum values, and the robustness of an estimation algorithm to noise is improved and the calculated amount of the iteration process is reduced by reducing the variables of the target function. 本发明