JVLDLoc: a Joint Optimization of Visual-LiDAR Constraints and Direction Priors for Localization in Driving Scenario

TL;DR

JVLDLoc introduces a novel joint optimization framework that fuses visual, LiDAR, and direction priors to improve localization accuracy in autonomous driving scenarios, reducing drift and enhancing robustness.

Contribution

The paper presents a new method that integrates direction priors with visual-LiDAR SLAM in a tightly-coupled optimization, improving localization accuracy over existing approaches.

Findings

Achieves lower localization error on KITTI datasets.

Reduces Absolute Pose Error compared to prior methods.

Validates effectiveness across multiple driving datasets.

Abstract

The ability for a moving agent to localize itself in environment is the basic demand for emerging applications, such as autonomous driving, etc. Many existing methods based on multiple sensors still suffer from drift. We propose a scheme that fuses map prior and vanishing points from images, which can establish an energy term that is only constrained on rotation, called the direction projection error. Then we embed these direction priors into a visual-LiDAR SLAM system that integrates camera and LiDAR measurements in a tightly-coupled way at backend. Specifically, our method generates visual reprojection error and point to Implicit Moving Least Square(IMLS) surface of scan constraints, and solves them jointly along with direction projection error at global optimization. Experiments on KITTI, KITTI-360 and Oxford Radar Robotcar show that we achieve lower localization error or Absolute Pose Error (APE) than prior map, which validates our method is effective.