AprilTag-Aided Pose Estimation Optimization of Landmark VSLAM

For visual Simultaneous Localization and Mapping (VSLAM), the camera positional estimation is deeply affected by environmental factors such as the surrounding feature situation and light condition. In this case, the accuracy and robustness of the positional estimation in visual odometry can be improved by the assistance of artificial landmarks. Based on this, a correction method based on AprilTag is proposed for ORB-SLAM3 to optimize the pose estimation of landmark VSLAM. Firstly, a point cloud map is built with Tag mapping embedded through the high accuracy 6 DOF positional transformation data provided by AprilTag. With this, the tracking module of ORB-SLAM3 is improved the positioning accuracy during pose estimating. In this way, the computational resources for estimating the camera pose can be reduced while guaranteeing the accuracy of pose estimation. Through subsequent validation experiments, the effectiveness of AprilTag-aided solution for ORB-SLAM3 is demonstrated for pose estimation optimization of landmark VSLAM.