Exploiting polar symmetry in designing equivariant observers for vision-based motion estimation
Accurately estimating camera motion from image sequences poses a significant challenge in computer vision and robotics. Many computer vision methods first compute the essential matrix associated with a motion and then extract orientation and normalized translation as inputs to pose estimation, reconstructing the scene scale (that is unobservable in the epipolar construction) from separate information. In this paper, we design a continuous-time filter that exploits the same perspective by using the epipolar constraint to define pseudo-measurements. We propose a novel polar symmetry on the pose of the camera that makes these measurements equivariant. This allows us to apply recent results from equivariant systems theory to estimating pose. We provide a novel explicit persistence of excitation condition to characterize observability of the full pose, ensuring reconstruction of the scale parameter that is not directly observable in the epipolar construction.
Code (0)
등록된 구현이 없습니다.
Tasks
Motion EstimationPose EstimationSimilar Papers 제목 키워드 기반
Observer Design for Nonlinear Systems with Equivariance
Equivariance is a common and natural property of many nonlinear control systems, especially those associated with models of mechatronic and navigation systems. Such systems admit a symmetry, associated with the equivaria…
Symmetry-Aware Steering of Equivariant Diffusion Policies: Benefits and Limits
Equivariant diffusion policies (EDPs) combine the generative expressivity of diffusion models with the strong generalization and sample efficiency afforded by geometric symmetries. While steering these policies with rein…
Reinforcement LearningLearning to Discover Reflection Symmetry via Polar Matching Convolution
The task of reflection symmetry detection remains challenging due to significant variations and ambiguities of symmetry patterns in the wild. Furthermore, since the local regions are required to match in reflection for d…
Self-Supervised LearningSymmetry DetectionEqR: Equivariant Representations for Data-Efficient Reinforcement Learning
We study different notions of equivariance as an inductive bias in Reinforcement Learning (RL) and propose new mechanisms for recovering representations that are equivariant to both an agent’s action, and symmetry transf…
Atari GamesInductive Biasreinforcement-learningReinforcement Learning+1Equivariant Filters are Equivariant
Observers for systems with Lie group symmetries are an active area of research that is seeing significant impact in a number of practical domains, including aerospace, robotics, and mechatronics. This paper builds on the…