Virtual-point-based Solutions to Handle Generalized Absolute Pose Problem

2026-06-08

AI Summary

This work addresses the challenge that existing Perspective-n-Point (PnP) solvers struggle with generalized absolute pose problems involving multiple projection centers. To bridge this gap, the authors propose a virtual point modeling approach that unifies standard PnP and generalized pose estimation within a single framework, enabling conventional PnP solvers to be extended to the generalized setting. Building upon this formulation, they derive three novel solvers based on distinct rotation parametrizations—Cayley parameters, quaternions, and rotation matrices—each balancing accuracy and computational efficiency while preserving global optimality. Experimental results demonstrate that the proposed methods significantly outperform current state-of-the-art generalized pose solvers under heteroscedastic noise conditions.

Abstract

Multi-camera systems are increasingly adopted in robotics and autonomous navigation for their wide field of view, flexibility, and fault tolerance. Nevertheless, existing PnP solvers fail to handle multiple projection centers. This paper introduces a virtual point formulation that bridges the standard PnP and generalized pose problems, enabling a unified pipeline that transforms existing PnP solvers into generalized pose solvers. Based on this framework, we derive three Virtual-point-based Generalized Pose solvers, namely VGPc, VGPq, and VGPr, leveraging Cayley, quaternion, and rotation-matrix parameterizations, respectively. Extensive experiments demonstrate that the proposed solvers inherit the accuracy and efficiency of original PnP algorithms while significantly outperforming existing generalized solvers. Specifically, VGPc achieves higher estimation accuracy under heteroscedastic noise conditions, VGPq maintains global optimality, whereas VGPr provides superior computational efficiency without accuracy degradation.

Problem

• generalized absolute pose
• multi-camera systems
• PnP solvers
• multiple projection centers

Innovation

• virtual point
• generalized pose estimation
• multi-camera systems
• PnP solvers
• heteroscedastic noise