GODSAC*: Graph Optimized DSAC* for Robot Relocalization
This addresses robot relocalization in outdoor environments, offering a specific improvement over existing methods.
The paper tackles the problem of camera pose estimation in large outdoor areas where existing deep learning methods perform poorly due to feature scarcity, and it proposes GODSAC*, which augments neural network predictions with odometry data via pose graph optimization, achieving improved accuracy over state-of-the-art approaches.
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.