Globally-Optimal Inlier Maximization for Relative Pose Estimation Under Planar Motion.

Haotian Liu,Alois Knoll,Ruiqi Zhang,Yinlong Liu,Zichen Liang,Guang Chen

doi:10.3389/fnbot.2022.820703

Abstract

Planar motion constraint occurs in visual odometry (VO) and SLAM for Automated Guided Vehicles (AGVs) or mobile robots in general. Conventionally, two-point solvers can be nested to RANdom SAmple Consensus to reject outliers in real data, but the performance descends when the ratio of outliers goes high. This study proposes a globally-optimal Branch-and-Bound (BnB) solver for relative pose estimation under general planar motion, which aims to figure out the globally-optimal solution even under a quite noisy environment. Through reasonable modification of the motion equation, we decouple the relative pose into relative rotation and translation so that a simplified bounding strategy can be applied. It enhances the efficiency of the BnB technique. Experimental results support the global optimality and demonstrate that the proposed method performs more robustly than existing approaches. In addition, the proposed algorithm outperforms state-of-art methods in global optimality under the varying level of outliers.

Highlights

Last decades witness the rapid development of frame to frame relative pose estimation in the field of computer vision, especially in visual odometry (VO), SLAM (Mur-Artal et al, 2015; Mur-Artal and Tardós, 2017), structure-from-motion (Schonberger and Frahm, 2016), 3D action understanding (Chen et al, 2014, 2015), trajectory online adaption (Luo et al, 2020, 2021) and gesture recognition (Qi and Aliverti, 2019; Qi et al, 2021)
We propose a globally-optimal BnB algorithm for the relative pose problem under planar motion constraint, where the algorithm is suitable for mobile robots or Automated Guided Vehicles (AGVs)
Recent studies on relative pose estimation are targeted at more robust and faster methods, which will improve the performance of AGVs and robots

Summary

Introduction

Last decades witness the rapid development of frame to frame relative pose estimation in the field of computer vision, especially in visual odometry (VO), SLAM (Mur-Artal et al, 2015; Mur-Artal and Tardós, 2017), structure-from-motion (Schonberger and Frahm, 2016), 3D action understanding (Chen et al, 2014, 2015), trajectory online adaption (Luo et al, 2020, 2021) and gesture recognition (Qi and Aliverti, 2019; Qi et al, 2021). Relative pose estimation solvers recover correct relative 3D rotation and translation of the camera based on feature matching of consecutive image pairs to support the mentioned above applications, which promotes the mutual development of pose estimation, AGVs, and mobile robotic technology. We focus on tackling the problem under planar motion constraint, e.g., the on-road vehicle is equipped with a forward looking camera. Such kinematic constraint is quite common and practical for Automated Guided Vehicles (AGVs) and robots designed for many real applications

Objectives

Methods

Conclusion