Virtual-point-based Solutions to Handle Generalized Absolute Pose Problem
2026-06-08 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors address a problem in robotics where current camera pose estimation methods cannot handle multiple camera centers well. They propose a new way to connect standard pose-solving techniques with more complex setups by introducing a 'virtual point' concept. Using this idea, they create three new solvers that adapt existing methods to work better with multi-camera systems. Their experiments show these new solvers keep the accuracy and speed of old methods but work better in challenging conditions. Each solver has specific strengths: accuracy with varied noise, guaranteed best solutions, or faster computation.
PnP (Perspective-n-Point)generalized pose problemcamera pose estimationmulti-camera systemsCayley parameterizationquaternionrotation matrixheteroscedastic noiseroboticsautonomous navigation
Authors
Bin Li, Banglei Guan, Shunkun Liang, Yang Shang
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.