CoordiNet: uncertainty-aware pose regressor for reliable vehicle localization

TL;DR

CoordiNet is a CNN-based vehicle localization method that estimates camera pose and uncertainty from a single image, outperforming previous methods on large-scale urban datasets and enabling real-time deployment.

Contribution

The paper introduces CoordiNet, a novel fully convolutional architecture that embeds scene geometry for improved pose estimation and uncertainty prediction in vehicle localization.

Findings

Outperforms previous methods on Oxford RobotCar dataset with 8m error

Achieves 29cm median error in real-time urban vehicle localization

Provides reliable uncertainty estimates for pose predictions

Abstract

In this paper, we investigate visual-based camera re-localization with neural networks for robotics and autonomous vehicles applications. Our solution is a CNN-based algorithm which predicts camera pose (3D translation and 3D rotation) directly from a single image. It also provides an uncertainty estimate of the pose. Pose and uncertainty are learned together with a single loss function and are fused at test time with an EKF. Furthermore, we propose a new fully convolutional architecture, named CoordiNet, designed to embed some of the scene geometry. Our framework outperforms comparable methods on the largest available benchmark, the Oxford RobotCar dataset, with an average error of 8 meters where previous best was 19 meters. We have also investigated the performance of our method on large scenes for real time (18 fps) vehicle localization. In this setup, structure-based methods require a large database, and we show that our proposal is a reliable alternative, achieving 29cm median error in a 1.9km loop in a busy urban area