Enhancing Real-Time Visual SLAM with Distant Landmarks in Large-Scale Environments
The efficacy of visual Simultaneous Localization and Mapping (SLAM) diminishes in large-scale environments due to challenges in identifying distant landmarks, leading to a limited perception range and trajectory drift.This paper presents a practical method to enhance the accuracy of feature-based real-time visual SLAM for compact unmanned vehicles