2-Entity Random Sample Consensus for Robust Visual Localization: Framework, Methods, and Verifications

Robust and efficient visual localization is essential for numerous robotic applications. However, it remains a challenging problem especially when significant environmental or perspective changes are present, as there high percentage of outliers, i.e., incorrect feature matches between the query image map. In this article, we propose novel 2-entity random sample consensus (RANSAC) framework using three-dimensional-two-dimensional point line with aid inertial measurements derive minimal closed-form solutions only 1 2 both monocular multi-camera system. The proposed RANSAC can achieve higher robustness against outliers multiple types features utilized number needed to compute pose reduced. Furthermore, learning-based sampling strategy selection mechanism scoring network be adaptive different characteristics such structured unstructured. Finally, simulation real-world experiments performed validate effectiveness method in scenarios long-term changes.