GODSAC*: Graph Optimized DSAC* for Robot Relocalization

TL;DR

GODSAC* is a novel graph-optimized approach that enhances neural network-based camera pose estimation by integrating noisy odometry, significantly improving accuracy in large outdoor environments for visual SLAM.

Contribution

It introduces GODSAC*, a method combining neural pose predictions with pose graph optimization to address outdoor localization challenges.

Findings

GODSAC* outperforms existing methods in pose accuracy.

The approach effectively integrates odometry data.

Experimental results demonstrate improved outdoor localization.

Abstract

Deep learning based camera pose estimation from monocular camera images has seen a recent uptake in Visual SLAM research. Even though such pose estimation approaches have excellent results in small confined areas like offices and apartment buildings, they tend to do poorly when applied to larger areas like outdoor settings, mainly because of the scarcity of distinctive features. We propose GODSAC* as a camera pose estimation approach that augments pose predictions from a trained neural network with noisy odometry data through the optimization of a pose graph. GODSAC* outperforms the state-of-the-art approaches in pose estimation accuracy, as we demonstrate in our experiments.